remyelination alters the pattern of myelin in the cerebral cortex · cerebral cortex, we performed...

1

Remyelination alters the pattern of myelin in the cerebral cortex 1 2 Jennifer Orthmann-Murphy1,2*, Cody L. Call1*, Gian Carlo Molina-Castro1, Yu Chen Hsieh1¥, 3 Matthew N. Rasband3, Peter A. Calabresi4, Dwight E. Bergles1,5 4 5 6 1The Solomon Snyder Department of Neuroscience, Johns Hopkins University, Baltimore, 7 Maryland 21205, USA 8 9 2Department of Neurology, Perelman School of Medicine, University of Pennsylvania, 10 Philadelphia, Pennsylvania 19104, USA 11 12 3Department of Neuroscience, Baylor College of Medicine, One Baylor Plaza, Houston, TX 13 77030, USA 14 15 4Department of Neurology Johns Hopkins University, Baltimore, Maryland 21205, USA 16 17 5Johns Hopkins University Kavli Neuroscience Discovery Institute, Baltimore, Maryland 21205 18 USA 19 20 21 * Equal contribution 22 23 ¥Current address: 24 Department of Molecular Biology, Simches Research Center, Massachusetts General Hospital, 25 185 Cambridge Street, Boston, MA 02114 26 27 28 Correspondence: 29 Dwight E. Bergles, PhD 30 Solomon H. Snyder Department of Neuroscience 31 Johns Hopkins University School of Medicine 32 725 N. Wolfe Street, Baltimore, MD 21205, USA 33 P:410-955-6939 34 [email protected] 35

.CC-BY 4.0 International license(which was not certified by peer review) is the author/funder. It is made available under aThe copyright holder for this preprintthis version posted March 9, 2020. . https://doi.org/10.1101/2020.03.09.983569doi: bioRxiv preprint

https://doi.org/10.1101/2020.03.09.983569

http://creativecommons.org/licenses/by/4.0/

2

ABSTRACT 36 37 Destruction of oligodendrocytes and myelin sheaths in cortical gray matter profoundly alters 38 neural activity and is associated with cognitive disability in multiple sclerosis (MS). Myelin can 39 be restored by regenerating oligodendrocytes from resident progenitors; however, it is not 40 known whether regeneration restores the complex myelination patterns in cortical circuits. Here 41 we performed time lapse in vivo two photon imaging in somatosensory cortex of adult mice to 42 define the kinetics and specificity of myelin regeneration after acute oligodendrocyte ablation. 43 These longitudinal studies revealed that the pattern of myelination in cortex changed 44 dramatically after regeneration, as new oligodendrocytes were formed in different locations and 45 new sheaths were often established along axon segments previously lacking myelin. Despite 46 the dramatic increase in axonal territory available, oligodendrogenesis was persistently impaired 47 in deeper cortical layers that experienced higher gliosis. The repeated reorganization of myelin 48 patterns in MS may alter circuit function and contribute to cognitive decline. 49 50 51 INTRODUCTION 52 53 Oligodendrocytes form concentric sheets of membrane around axons that enhance the speed of 54 action potential propagation, provide metabolic support and control neuronal excitability through 55 ion homeostasis. Consequently, loss of oligodendrocytes and myelin can alter the firing 56 behavior of neurons and impair their survival, leading to profound disability in diseases such as 57 multiple sclerosis (MS), in which the immune system inappropriately targets myelin for 58 destruction. In both relapsing-remitting forms of MS and the cuprizone model of demyelination in 59 mouse, the CNS is capable of spontaneous remyelination through mobilization of 60 oligodendrocyte precursor cells (OPCs) (Baxi et al., 2017; Chang, Nishiyama, Peterson, 61 Prineas, & Trapp, 2000; Chang et al., 2012), which remain abundant in both gray and white 62 matter throughout adulthood (Dimou, Simon, Kirchhoff, Takebayashi, & Götz, 2008; Hughes, 63 Kang, Fukaya, & Bergles, 2013; Young et al., 2013). The highly ordered structure of myelin in 64 white matter tracts has enabled in vivo longitudinal tracking of inflammatory demyelinating 65 lesions using magnetic resonance imaging (MRI); however, due to the low spatial resolution of 66 standard MRI sequences (Oh et al., 2019) and the indirect nature of MR methods used to 67 assess the integrity of myelin (Walhovd, Johansen-Berg, & Káradóttir, 2014), our knowledge 68


https://doi.org/10.1101/2020.03.09.983569


3

about the dynamics of OPC recruitment, oligodendrogenesis and remyelination within specific 69 neural circuits remains limited. 70 In vivo studies of remyelination have focused primarily on white matter, where assessments 71 of myelin are aided by the high density and symmetrical alignment of myelin sheaths; however, 72 postmortem histological analysis (Kidd et al., 1999; Kutzelnigg, 2005; Lucchinetti et al., 2011; 73 Peterson, Bö, Mörk, Chang, & Trapp, 2001) and new in vivo MRI and PET imaging methods 74 (Beck et al., 2018; Filippi et al., 2014; Herranz et al., 2019; Kilsdonk et al., 2013; R. Magliozzi, 75 Reynolds, & Calabrese, 2018) indicate that demyelination is also prevalent in cortical gray 76 matter of MS patients. Cortical lesion load correlates with signs of physical and cognitive 77 disability, including cognitive impairment, fatigue and memory loss (Calabrese et al., 2012; 78 Nielsen et al., 2013). Nevertheless, much less is known about the capacity for repair of myelin in 79 cortical circuits. Defining how gray matter lesions are resolved in vivo is critical for both MS 80 prognosis and the development of new therapies to promote remyelination. 81 Unlike white matter, myelination patterns in the cerebral cortex are highly variable, with 82 sheath content varying between cortical regions, among different types of neurons and even 83 along the length of individual axons (Micheva et al., 2016; Tomassy et al., 2014). Despite this 84 evidence of discontinuous myelination, recent in vivo imaging studies indicate that both 85 oligodendrocytes and individual myelin sheaths are remarkably stable in the adult brain (Hill, Li, 86 & Grutzendler, 2018; Hughes, Orthmann-Murphy, Langseth, & Bergles, 2018), suggesting that 87 maintaining precise sheath placement is important for cortical function. However, the complex 88 arrangement of cortical myelin presents significant challenges for repair and it is unknown 89 whether precise myelination patterns in the cortex are restored following a demyelinating event. 90 In vivo two photon fluorescence microscopy allows visualization of oligodendrogenesis and 91 myelin sheath formation in mammalian circuits at high resolution, providing the means to define 92 both the dynamics and specificity of regeneration (Hughes et al., 2018). Here, we used this 93 high-resolution imaging method to define the extent of oligodendrocyte regeneration and the 94 specificity of myelin replacement within the adult somatosensory cortex after demyelination. 95 Unexpectedly, our studies indicate that oligodendrocytes are regenerated in locations distinct 96 from those occupied before injury. Despite the additional available axonal territory for 97 myelination, regenerated oligodendrocytes had a similar size and structure; as a result, only a 98 fraction of prior sheaths were replaced and many new sheaths were formed on previously 99 unmyelinated regions of axons. Conversely, in regions of high territory overlap, new 100 oligodendrocytes often formed sheaths along the same segment of specific axons, 101 demonstrating that precise repair is possible. Together, these in vivo findings indicate that 102


https://doi.org/10.1101/2020.03.09.983569


4

regeneration of oligodendrocytes in the cortex creates a new pattern of myelination, with 103 important implications for the restoration of sensory processing and cognition. 104 105 RESULTS 106 107 Inefficient regeneration of oligodendrocytes in cortical gray matter 108 To define the dynamics of oligodendrocyte regeneration and axonal remyelination in the 109 cerebral cortex, we performed longitudinal two photon imaging through a cranial window placed 110 over the barrel field of the somatosensory cortex in transgenic mice that express EGFP under 111 control of the promoter/enhancer for myelin-associated oligodendrocyte basic protein (Mobp-112 EGFP) mice (Hughes et al., 2018) (Figure 1A). In these mice, complete oligodendrocyte 113 morphologies could be resolved in vivo, including cytoplasmic processes and individual myelin 114 internodes within the upper layers of the cortex. In these regions, oligodendrocytes ensheath a 115 select group of axons, including long-range axonal projections oriented horizontally to the pia 116 (Bock et al., 2011) (Figure 1B), and in deeper layers, vertically-oriented axons belonging to both 117 local cortical neurons and long-range projections (Figure 1A,C). To induce demyelination, young 118 adult Mobp-EGFP mice (age 8-12 weeks) were fed chow mixed with 0.2% cuprizone, a copper 119 chelator that induces robust fragmentation and apoptosis of oligodendrocytes (Vega-Riquer, 120 Mendez-Victoriano, Morales-Luckie, & Gonzalez-Perez, 2019) (Supplementary Figure 1), and 121 multiple volumes (425 μm x 425 μm x 550 μm) corresponding to layers I–IV were imaged 122 repeatedly prior to injury, during demyelination and through recovery for up to 12 weeks (Figure 123 1D, Supplementary Video 1). 124 Oligodendrocytes and individual myelin sheaths are extraordinarily stable in the adult 125 brain (Hill et al., 2018; Hughes et al., 2018; Yeung et al., 2014); however, the amount of myelin 126 within these circuits is not static, as new oligodendrocytes continue to be generated in the adult 127 CNS, each of which produces dozens of sheaths (Hughes et al., 2018; Kang, Fukaya, Yang, 128 Rothstein, & Bergles, 2010; Xiao et al., 2016). This phenomenon was visible during in vivo 129 imaging in Mobp-EGFP mice as new EGFP-expressing (EGFP+) oligodendrocytes appeared 130 within the imaging field (Figure 1E, F, H), continuously adding to the baseline oligodendrocyte 131 population (Figure 1H; Supplementary Video 2). When mice were fed cuprizone for three 132 weeks, > 90% of the baseline population of oligodendrocytes within the upper layers of cortex 133 degenerated (94.2 ± 0.05%; N = 6 mice, mean ± SEM) (Figure 1E,G,I; Supplementary Video 3), 134 with most cell loss occurring after cessation of cuprizone exposure (Figure 1I,J). New 135 oligodendrocytes initially appeared rapidly during the recovery phase (Figure 1K); however, this 136


https://doi.org/10.1101/2020.03.09.983569


5

burst of oligodendrogenesis was not sustained, and as a consequence, only about half of the 137 oligodendrocytes (55.2 ± 0.03%) were replaced after nine weeks of recovery. Extrapolating from 138 the rate of addition from the last recovery time-point (3.5 ± 0.5% per week, Figure 1K), it would 139 take approximately three additional months (~12.8 weeks) to achieve the density of 140 oligodendrocytes prior to cuprizone, which is ~two-fold greater in middle-aged cortices (Hughes 141 et al., 2018). Thus, these young mice would be > 13 months old by the time they achieved full 142 recovery, assuming a constant rate of addition. However, as oligodendrogenesis declines with 143 age in both control and cuprizone-treated mice (Figure 1K) relative to age-matched control 144 mice, cuprizone-treated mice may never reach a normal oligodendrocyte density after 145 demyelination. These results indicate that although a prominent regenerative response is 146 initiated in cortical gray matter, oligodendrocyte regeneration is much slower (Baxi et al., 2017) 147 and less complete than in white matter (Baxi et al., 2017; Gudi et al., 2009; Jeffery & 148 Blakemore, 1995; Matsushima & Morell, 2001). 149 150 Layer specific differences in cortical remyelination 151 The cerebral cortex is a highly layered structure, in which genetically and morphologically 152 distinct neurons form specialized subnetworks (Lodato & Arlotta, 2015), raising the possibility 153 that they may adopt different myelination patterns to optimize circuit capabilities (Micheva et al., 154 2016; Stedehouder et al., 2017; Tomassy et al., 2014). Indeed, myelination patterns are highly 155 non-uniform across the cortex (Supplementary Figure 2) and both oligodendrocyte density and 156 myelin content increase with depth from the cortical surface (Hughes et al., 2018) (Figure 1C). 157 However, it is not known whether the capacity for myelin repair is comparable between cortical 158 layers. To examine depth-dependent changes in oligodendrogenesis, we subdivided imaging 159 volumes into three 100 μm zones starting from the pial surface (Figure 2A-C). In control mice, 160 the proportional rates of oligodendrocyte addition (top: 2.62 ± 0.37%; bottom: 2.93 ± 0.26% per 161 week) were similar between the top zone (0-100 μm) and bottom zone (200-300 μm), despite 162 their dramatically different oligodendrocyte densities (0-100 μm: 17 ± 2 cells; 200-300 μm: 79 ± 163 10 cells, N = 11 mice @ baseline) (Figure 2D,E); the only exception was for times when no cells 164 were incorporated into the top zone (Figure 2E, p = 0.0036 @ 5 weeks, N-way ANOVA with 165 Bonferroni correction for multiple comparisons). These findings suggest that oligodendrocyte 166 enrichment in deeper layers occurs early in development, but then proceeds at similar rates 167 across cortical layers in adulthood. 168

In an efficient regenerative process, cell generation would be matched to cell loss; 169 however, we found that oligodendrocyte regeneration was not proportional to their original 170


https://doi.org/10.1101/2020.03.09.983569


6

density. Oligodendrocyte recovery was highly efficient in the top 100 μm zone, reaching 85.2 ± 171 0.17% of the baseline oligodendrocyte population nine weeks after cuprizone, but only 55.5 ± 172 0.11% of baseline after nine weeks in the bottom 100 μm zone (Figure 2C). The peak of this 173 oligodendrocyte loss and replacement occurred during the first two weeks of recovery post-174 cuprizone (Figure 1J,K, Figure 2F,G). During this period, there was both a proportionally higher 175 rate of cell loss and lower rate of cell replacement in the bottom 100 μm zone compared to the 176 upper zone (Figure 2F,G) (loss @ week 4, p = 0.044; addition @ week 4, p = 0.036, N-way 177 ANOVA with Bonferroni correction for multiple comparisons), indicating that the ability to 178 balance cell loss with replacement declines with depth during this initial period. Although 179 proportionally lower, more oligodendrocytes were generated per week in deeper layers (Figure 180 2H, green bars) (0-100 μm: 0.5-3.3 cells/week; 200-300 μm: 2.8 – 7.7 cells/week, p = 5.82 x 10-4 181 @ week 4, p = 0.046 @ week 5, N-way ANOVA with Bonferroni correction for multiple 182 comparisons), but this enhanced oligodendrogenesis was not sufficient to compensate for the 183 greater cell loss (Figure 2H, magenta bars). This relative lag in regeneration is clearly visible in 184 maps indicating where newly generated oligodendrocytes were formed with regard to depth and 185 their corresponding density histograms (Figure 2I, green circles and bars). This analysis 186 highlights that the absolute number of oligodendrocytes integrated was remarkably similar in the 187 top and bottom zones during the first few weeks of recovery, suggesting that there are factors 188 that suppress regeneration of needed oligodendrocytes in deeper cortical layers. 189 A decrease in the pool of progenitors or formation of reactive astrocytes are potential 190 candidates to impair oligodendrocyte regeneration in deeper cortical layers. Although OPCs are 191 slightly less abundant in the deeper layers compared to the surface 100 μm (Supplementary 192 Figure 3A,C,D) (p = 2.084 x 10–13, N-way ANOVA, by depth), there was no evidence of 193 persistent OPC depletion during recovery (Supplementary Figure 3D,F) (p = 0.086, Kruskal-194 Wallis one-way ANOVA). Reactive astrocytes, a known pathological feature of both the 195 cuprizone model and cortical demyelinating lesions (Chang et al., 2012; Skripuletz et al., 2008), 196 can impair OPC differentiation by secreting cytokines (Kirby et al., 2019; Su et al., 2011; Zhang 197 et al., 2010), but whether reactive astrocytes impair recovery differently as a function of cortical 198 depth is unknown. GFAP+ astrocytes are relatively rare in the deeper (200-500 μm) regions of 199 cortex in naïve mice; however, following cuprizone-treatment, their number increased nearly 200 ten-fold (p = 5.86 x 10-16, N-way ANOVA, by time-point). This enhanced GFAP expression 201 remained elevated throughout the recovery period (Supplementary Figure 3G) (p = 0.006, 202 Kruskal-Wallis one-way ANOVA, with Fisher’s correction for multiple comparisons) and they 203 retained a reactive morphology, even after five weeks of recovery (Supplementary Figure 204


https://doi.org/10.1101/2020.03.09.983569


7

3A,B,E,G). In contrast, astrocytes in the surface 100 μm, while exhibiting higher GFAP 205 immunoreactivity at baseline, exhibited only a transient increase that was not sustained past two 206 weeks of recovery (Supplementary Figure 3E). These findings highlight that the gliotic response 207 to demyelination varies in magnitude and duration across the cortex, which may impair the 208 recovery of gray matter regions with higher myelin content. 209 210 Regeneration alters the pattern of myelination in the cortex 211 Myelination patterns are distinct among different neuron classes in the cortex (Micheva et al., 212 2016; Stedehouder et al., 2017; Tomassy et al., 2014), and can be discontinuous, in which 213 myelin segments along individual axons are separated by long regions of bare axon (Tomassy 214 et al., 2014). Even these isolated myelin segments are highly stable (Hill et al., 2018; Hughes et 215 al., 2018), suggesting that preservation of these patterns is important to support cortical 216 function, and therefore that recreation of these patterns should be a goal of the regenerative 217 process. We hypothesized that given the sparseness of myelination in the cortex, new 218 oligodendrocytes generated after demyelination should be formed in locations very close to the 219 original population. To explore the spatial aspects of oligodendrocyte replacement, we mapped 220 the distribution of oligodendrocytes within 425 μm x 425 μm x 300 μm volumes in 3-D prior to 221 and after recovery from cuprizone, as well as the distribution of new cells generated in control 222 mice (Figure 3A,B, green circles). Unexpectedly, these maps revealed that the cell bodies of 223 regenerated oligodendrocytes occupied positions distinct from the original cells (see also Figure 224 1E). To quantify displacement of newly formed oligodendrocytes from original locations, we 225 determined the Euclidean distance between nearest neighbors (nearest neighbor distance, 226 NND) of every oligodendrocyte in this region at baseline and at five weeks of recovery (or eight 227 weeks in control). Small changes in cell position were observed for oligodendrocytes in control 228 mice (Figure 3C, Stable cells); however, addition of new oligodendrocytes to the cortex over this 229 period did not significantly alter NND (All cells vs. Stable cells, 0-300 μm: p = 0.284, N-way 230 ANOVA with Bonferroni correction for multiple comparisons). In contrast, regenerated 231 oligodendrocytes added to the cortex of cuprizone-treated mice were significantly displaced 232 relative to original locations (Figure 3C, All Cells vs. Stable cells) (p = 8.95 x 10-7, N-way 233 ANOVA with Bonferroni correction for multiple comparisons). This apparent rearrangement of 234 oligodendrocyte position was not due to differences in image quality or registration across the 235 time series, or to changes in the structure of the tissue due to cuprizone exposure, as 236 oligodendrocytes that survived in cuprizone exhibited the same average displacement as cells 237 in control mice over the course of eight weeks of imaging (Figure 3C, Stable cells, Control vs. 238


https://doi.org/10.1101/2020.03.09.983569


8

Regenerated, p > 0.05, N-way ANOVA with Bonferroni correction for multiple comparisons). 239 This increase in NND was also not due to incomplete recovery from demyelination, as NND in 240 the upper zone (0-100 μm) was greater than the average NDD across the whole volume (0-300 241 μm), despite the proportionally greater regeneration in layer I (Figure 2C). These results 242 demonstrate that regenerated oligodendrocytes occupy locations within the parenchyma that 243 are distinct from those present at baseline, and therefore may establish a new pattern of 244 myelination. 245 The new locations of oligodendrocytes after a demyelinating event may not preclude 246 these cells from myelinating the same axonal segments, as cortical oligodendrocytes can 247 extend long cytoplasmic processes to form sheaths distant from the cell body (Chong et al., 248 2012; Murtie, Macklin, & Corfas, 2007). To determine whether regenerated oligodendrocytes 249 restore the pattern of myelination by extending longer processes, we reconstructed their 250 complete morphologies and compared these to oligodendrocytes generated at the same age in 251 control mice. For new oligodendrocytes that appeared in layer I in control (10 cells, N = 3 mice) 252 or cuprizone-treated mice (9 cells, N = 3 mice), each process extending from the cell body and 253 every myelin sheath connected to these processes were traced when the cells first appeared in 254 the imaging volume (Figure 3D), and for every 2 - 3 days for up to 14 days. In both control and 255 cuprizone treated mice, newly generated oligodendrocytes exhibited an initial period of 256 refinement after first appearance (governed by the onset of Mobp promoter activity and EGFP 257 accumulation in the cytoplasm), characterized by sheath extensions and retractions 258 (Supplementary Fig. 4A-D), pruning of entire myelin sheaths and removal of cytoplasmic 259 processes (Supplementary Fig. 4E-H), before reaching a stable morphology (Supplementary 260 Fig. 4I,J). The final position of the cytoplasmic process along the length of the sheaths (the likely 261 point of sheath initiation) was randomly distributed along the length of the internode 262 (Supplementary Fig. 4K,L). Notably, the initial dynamics of these newly formed oligodendrocytes 263 are remarkably similar to those described in the developing zebrafish spinal cord (Auer, 264 Vagionitis, & Czopka, 2018; Czopka, Ffrench-Constant, & Lyons, 2013), suggesting that the 265 developmental sequence of oligodendrocytes is both highly conserved and cell intrinsic. 266 This time-lapse analysis revealed that the morphological plasticity of newly formed 267 oligodendrocytes lasts for more than one week in the cortex (Supplementary Figure 4I,J); 268 therefore, comparisons between control and regenerated oligodendrocytes were made from 269 cells 12-14 days after first appearance when they reached their mature form. Although 270 regenerated oligodendrocytes had access to much more axonal territory, they produced similar 271 numbers of sheaths (Figure 3E) (Control: 53 ± 3 sheaths, 10 cells, N = 3 mice; Regenerated: 51 272


https://doi.org/10.1101/2020.03.09.983569


9

± 2 sheaths, 9 cells, N = 3 mice, p = 0.628, unpaired two-tailed t-test). However, regenerated 273 cells formed more total myelin (Figure 3F) (Total sheath length: Control, 3.17 ± 0.16 mm; 274 Regenerated: 3.80 ± 0.23 mm, p = 0.041, unpaired two-tailed t-test) by extending slightly longer 275 sheaths (Figure 3G) (average sheath length: Control, 62.6 ± 2.6 μm; Regenerated, 72.3 ± 2.2 276 μm, p = 0.012, unpaired two-tailed t-test), despite having similar distributions of sheath lengths 277 (Figure 3H). 278 If regenerated oligodendrocytes reach the same axonal regions from a greater distance 279 away, their processes should be more polarized; however, 2-D maps of sheaths arising from 280 single cells, revealed that they exhibited a similar radial morphology (Figure 3D). To obtain a 281 quantitative measure of polarization, vectors were calculated from the cell body to each 282 paranode (Figure 3I) and mean radial histograms for new and remyelinating cells were 283 calculated (Figure 3J). The average extent of polarization for control and regenerated cells was 284 not significantly different from uniformly radial (Control, –0.047 ± 1.30 (std) rad, p = 0.400; 285 Regenerated, 0.076 ± 1.30 (std) rad, p = 0.256, Hodges-Ajne test of non-uniformity) and 286 sheaths of new cells in control and those regenerated after cuprizone exhibited similar degrees 287 of circularity (p > 0.1, k = 462, Kuiper two-sample test). Thus, regenerated oligodendrocytes 288 formed in an environment with greater myelination targets have morphologies remarkably 289 similar to cells added to existing myelinated networks in naïve mice, suggesting that 290 oligodendrocyte structure is shaped by strong cell intrinsic mechanisms. 291 To estimate the extent of myelin sheath recovery in the somatosensory cortex, we 292 measured the overlap in cell territory between baseline and regenerated cells. Territories of 293 individual oligodendrocytes which existed at baseline, those generated in control, and those 294 regenerated following cuprizone were estimated using an ellipsoid centered at the center of 295 mass of the sheath arbor with the smallest radii in the x-y (oriented parallel to the pia) and z 296 planes (oriented perpendicular to pia) that encompassed at least 80% of the total sheath length. 297 These radii were averaged across all cells per condition to generate model ellipsoids (Figure 298 4A), representing the volume available to be myelinated by an individual oligodendrocyte. As 299 expected from the slightly longer myelin sheaths produced by remyelinating cells, their average 300 cell territory was significantly larger than control cells (Figure 4B) (x-y; Control: 75.7 ± 2.1 μm; 301 Regenerated: 85.2 ± 2.8 μm, p = 0.025, one-way ANOVA). Model ellipsoids were then centered 302 on the cell body coordinates in layer I from control and cuprizone-treated mice (examples shown 303 in Figure 4C,D) and their degree of territory overlap determined. This analysis revealed that 304 regenerated oligodendrocytes in cuprizone exposed mice exhibit only 59.1 ± 5.8 % territory 305 overlap (Range: 37.2 – 79.0%, N = 6) with oligodendrocytes that originally populated these 306


https://doi.org/10.1101/2020.03.09.983569


10

regions of the cortex. This convergence was slightly, but not significantly, higher than predicted 307 if the same number of regenerated cells were placed at random in the volume (44.3 ± 4.7%; p = 308 0.078 by t-test) (Figure 4E), suggesting that local factors may influence which progenitors 309 differentiate. Moreover, because regenerated cell territories were larger (Figure 4B), they 310 enclosed an average volume of 116 ± 14.1% of the total territory at baseline (Range: 87.4 – 311 181%) (Figure 4F). These data and those obtained from the NND analysis indicate that although 312 regenerated oligodendrocytes tend to be formed close to the sites of original cells, they 313 unexpectedly do not completely overlap with the baseline territory; thus, regenerated 314 oligodendrocytes are unlikely to access to the same complement of axons to myelinate, and 315 could potentially myelinate novel axon segments. 316 317 Regeneration of specific myelin segments 318 Although oligodendrocytes are formed in new locations during recovery from demyelination, 319 they have the opportunity to regenerate specific myelin sheaths in areas where they extend 320 processes into previously myelinated territories. To assess the extent of sheath replacement in 321 the somatosensory cortex, we acquired high resolution images in layer I in Mobp-EGFP mice, 322 allowing assessment of the position and length of individual myelin internodes. We randomly 323 selected 100 μm x 100 μm x 100 μm volumes in the image stacks (425 μm x 425 μm x 100 μm), 324 and traced the full length of internodes that entered the volume at baseline and at the five-week 325 recovery time point in control (N = 5) and cuprizone-treated mice (N = 5) (Figure 5A). Internodes 326 were classified as stable (supplied by the same cell at baseline and recovery), lost (present at 327 baseline and absent in recovery), replaced (present at baseline, lost and then replaced by a 328 new cell), or novel (not present at baseline, but present in recovery). Consistent with previous 329 findings (Hill et al., 2018; Hughes et al., 2018), in control mice almost all internodes (99.1 ± 0.5 330 %) present at baseline remained after five weeks (Figure 5A, gray processes), demonstrating 331 the high stability of myelin sheaths within the cortex. Generation of new oligodendrocytes within 332 this area led to the appearance of a substantial proportion (25.8 ± 8.7 %) of novel sheaths 333 (Figure 5A, green processes). In mice treated with cuprizone, 84.4 ± 2.7% of myelin sheaths 334 were destroyed in this region (Figure 5B), but sheath numbers were restored to baseline levels 335 after five weeks (Figure 5C). However, despite this apparent recovery of overall myelin content, 336 only 50.5 ± 2.9% of specific internodes were replaced, 32.4 ± 1.2% were lost (not replaced) and 337 47.6 ± 9.9 % novel internodes were formed (N = 5 mice) (Figure 5D), indicating that 338 regeneration leads to a dramatic change in the overall pattern of myelin within cortical circuits. 339


https://doi.org/10.1101/2020.03.09.983569


11

Analysis of single oligodendrocytes revealed that the degree of sheath replacement 340 varied considerably between regions. Oligodendrocytes regenerated in regions that originally 341 had a high density of sheaths devoted a larger proportion of their sheaths to replacement (R2 = 342 0.4232, 13 cells, N = 4 mice) (Figure 5E,F). This phenomenon was most evident in cells that 343 traversed boundaries of low and high myelin density. Nevertheless, rather than extending all 344 processes into the previously highly myelinated area, they retained a radial morphology (Fig. 345 3I); processes that extended into densely myelinated areas often formed sheaths on previously 346 myelinated axons, while processes that extended into the sparsely myelinated area typically 347 formed novel sheaths (Figure 5E,F). These results suggest that remyelination may be 348 opportunistic and based, at least initially, on chance interactions between processes and local 349 axon segments. This finding is consistent with the highly radial morphology of premyelinating 350 oligodendrocytes (Trapp, Nishiyama, Cheng, & Macklin, 1997), which would be optimized for 351 local search of the surrounding volume rather than directed, regional targeting of subsets of 352 axons. 353 Myelination patterns along axons in the cortex are highly variable, ranging from 354 continuous to sparsely myelinated in the same region (Hill et al., 2018; Hughes et al., 2018; 355 Micheva et al., 2016; Tomassy et al., 2014). To determine if there is a preference for replacing 356 specific types of sheaths during recovery from demyelination, we classified sheaths present at 357 baseline according to their neighbors: 0 neighbors within 5 μm (isolated), one node of Ranvier 358 (NOR) and one hemi-node (1 neighbor), or two NORs (2 neighbors) (Figure 6A,B; 359 Supplementary Videos 4,5). In control mice this pattern was highly stable, with similar 360 proportions of isolated and neighbored sheaths at baseline and eight weeks later (Figure 6C), a 361 pattern that we showed previously is stable through middle-age (Hughes et al., 2018). Indeed, 362 novel sheaths generated by new oligodendrocytes made roughly equal proportions of isolated 363 and neighbored sheaths (Figure 6D), indicating that new sheaths were added to previously-364 unmyelinated axons, as well as next to existing sheaths on myelinated axons, visible as higher 365 values in the upper right quadrant (relative to lower left quadrant) of the myelination matrix 366 (Figure 6E). We then assessed the extent of sheath replacement five weeks after the end of 367 cuprizone. Similar to that observed in control, oligodendrocytes in cuprizone-treated mice had 368 similar proportions of isolated and neighbored sheaths as they did at baseline (Figure 6F). While 369 lost (not replaced) and novel internodes exhibited both isolated and neighbored sheaths with 370 equal proportion, the majority of replaced internodes had at least one neighbor (79.5 ± 2.5 %) 371 (Figure 6G), visible in the bias towards higher values in the lower right quadrant in the 372 myelination matrix (Figure 6H). These data suggest that regenerated oligodendrocytes 373


https://doi.org/10.1101/2020.03.09.983569


12

preferentially formed sheaths on axons that were previously more heavily myelinated, 374 suggesting that the factors which promote greater myelination of specific axons is preserved 375 after demyelination. 376

Analysis of individual sheath position revealed the remarkable spatial specificity with 377 which sheaths were replaced. For both isolated sheaths and those along continuously 378 myelinated segments, sheaths were often formed close to the same position along axons, 379 resulting in NOR or hemi nodes at similar positions (Figure 7A, Supplementary Video 6). These 380 results suggest that there may be persistent landmarks along axons indicating the prior position 381 of nodes after myelin is removed (Figure 7B). To assess whether nodal components along 382 cortical axons remain after demyelination, we performed post-hoc immunostaining on tissue 383 from mice exposed to cuprizone for six weeks (longer than three weeks used above, to increase 384 the length of time that axons were devoid of myelin) with antibodies against Caspr, βIV-spectrin, 385 and Ankyrin-G, which together label paranodal and nodal regions along myelinated axons 386 (Susuki, Otani, & Rasband, 2016). High resolution z-stacks (135 μm x 135 μm x 30 μm, 2048 x 387 2048 pixels) of layer V-VI were acquired in order to capture a large population of nodes (Figure 388 7C,D), as the upper layers of cuprizone-treated mice had inadequate numbers for quantification 389 at this magnification. The full images were transformed into surfaces to allow automated 390 quantification (see Methods) (Figure 7E,F). βIV-spectrin puncta were classified as “nodes” if 391 they were flanked by Caspr+ puncta, “heminodes” if they were bound by only one side by 392 Caspr, and “isolated” if no flanking Caspr was visible (Figure 7B). As expected, the distribution 393 of nodes among these categories remained stable in control mice over two weeks (Figure 7G). 394 In contrast, the characteristics of nodes changed dramatically after exposure to cuprizone, with 395 a loss of nodes visible at four weeks, a time when ~50% of oligodendrocytes had degenerated 396 (Figures 1I, 7G). After six weeks in cuprizone, few Caspr+ puncta remained, and 397 correspondingly there were few nodes or heminodes (Figure 7F,G). However, even at this later 398 time point, many βIV-spectrin+ puncta were visible, with the greatest proportion of these 399 isolated from Caspr, suggesting that βIV-spectrin remains clustered for a prolonged period, 400 consistent with the hypothesis that many represent previous nodes (i.e. previously flanked by 401 Caspr). Thus, although prior studies suggest that NOR components are rapidly disassembled 402 and redistributed after demyelination (Coman et al., 2006; Craner et al., 2004; Dupree et al., 403 2004), these results indicate that βIV-spectrin, which links the underlying actin cytoskeleton to 404 integral membrane proteins within the node/paranode, remains clustered along axons, which 405 may provide the means to confine sheath extension to previously myelinated regions and allow 406 precise restoration of myelin sheath position during regeneration. 407


https://doi.org/10.1101/2020.03.09.983569


13

408 DISCUSSION 409 410 The organization of myelin in the cerebral cortex is remarkably diverse, with densities of myelin 411 sheaths varying between cortical regions, between axons from different classes of neurons and 412 even along individual axons within a given area. These patterns are established progressively 413 over many months, creating an extended developmental time course that results in pronounced 414 increases in myelin through adolescence and adulthood. Although progressive, 415 oligodendrogenesis and myelination can be enhanced by life experience and may be critical to 416 certain forms of learning (Gibson et al., 2014; Hughes et al., 2018; McKenzie et al., 2014). 417 However, once formed, oligodendrocytes and their complement of myelin sheaths are 418 extraordinarily stable, with cell survival and sheath position varying little over months and are 419 resistant to environmental changes such as sensory enrichment, in accordance with the high 420 stability of myelin proteins and the persistence of oligodendrocytes in the human CNS (Yeung et 421 al., 2014). Experience dependent changes in myelin appear to occur primarily through addition 422 of new sheaths arising from oligodendrogenesis (Hughes et al., 2018), but see also (Dutta et al., 423 2018; Gibson et al., 2014). The high stability of oligodendrocytes and their preferential 424 myelination of specific neurons, such as parvalbumin interneurons in layers II/III (Micheva et al., 425 2016; Stedehouder et al., 2017), suggest that preserving the overall pattern of myelin is 426 important to optimize and sustain the processing capabilities of these circuits. Consistent with 427 this hypothesis, demyelination within the cortex in diseases such as MS is closely associated 428 with cognitive impairment and increased morbidity. Moreover, many neurodegenerative 429 diseases and affective disorders are associated with profound alterations in myelin (Gouw et al., 430 2008; Ihara et al., 2010; Stedehouder & Kushner, 2017), and recent studies indicate that 431 demyelination triggers a dramatic decrease in excitatory synapses (Araújo et al., 2017; 432 Werneburg et al., 2020), suggesting that disruptions in these sheaths may also influence both 433 structural and functional aspects of circuits. Although preserving cortical myelination appears 434 paramount, the highly variable patterns of myelin within cortical circuits and the sparseness of 435 these neuron-glial associations create considerable challenges for regenerating specific myelin 436 sheaths after injury or disease. 437

Here, we used two photon in vivo imaging to examine the destruction and regeneration 438 of oligodendrocytes and myelin sheaths in cortical circuits of adult mice. These longitudinal, high 439 resolution studies revealed features of the regenerative process in the cortex that were not 440 previously known. First, oligodendrocytes were regenerated in new locations, yet had similar 441


https://doi.org/10.1101/2020.03.09.983569


14

morphologies. Second, regenerated oligodendrocytes often formed sheaths along portions of 442 axons that were previously unmyelinated, establishing a new pattern of myelination. Third, 443 oligodendrocyte regeneration was not uniform across the cortex and became less efficient with 444 depth from the cortical surface, in concert with the increasing density of myelin (prior to 445 oligodendrocyte destruction) and enhanced gliosis. Fourth, in areas of territory overlap, 446 regenerated oligodendrocytes were able to establish sheaths at similar positions along 447 previously myelinated axons, indicating that positional cues persist along axons long after 448 demyelination. Together, these findings reveal unexpected aspects of cortical remyelination, 449 raising new questions about the mechanisms that impair oligodendrocyte regeneration in 450 deeper cortical layers, the mechanisms that enable oligodendrocytes to identify and replace 451 individual myelin sheaths and the long-term consequences of circuit level changes in 452 myelination patterns. 453 454 Regenerative potential of cortical OPCs 455 In regenerative processes, cell loss and cell generation are typically closely coupled, helping to 456 ensure efficient replacement without energetically costly production of excess cells and further 457 tissue disruption (Biteau, Hochmuth, & Jasper, 2011). If this scenario were to prevail in the 458 CNS, the generation of new oligodendrocytes should be proportional to those lost, with much 459 higher oligodendrocyte production occurring in deeper layers. However, our results indicate that 460 oligodendrocyte replacement was remarkably constant across layers for many weeks after a 461 demyelinating event (Figure 2G-I), with regeneration in deeper layers lagging behind that 462 predicted for one-to-one replacement. This phenomenon could be explained if local 463 environmental factors in deeper layers suppress OPC differentiation. The larger number of 464 oligodendrocytes that degenerate in this area may inhibit regeneration by creating more 465 inflammation and myelin debris, factors known to suppress oligodendrogenesis. Consistent with 466 this hypothesis, reactive gliosis was more prominent and more prolonged in deeper layers of 467 cortex (Supplementary Figure 3A,B,E,G). Nevertheless, even after extended recovery (> 9 468 weeks), oligodendrocyte density remained much lower in these regions than present at 469 baseline. This regional suppression of oligodendrogenesis is particularly apparent when 470 assessing the rate of cell addition, as the initial burst in oligodendrogenesis that occurs just 471 following cell loss is not sustained despite the remaining cell deficit (Figure 2G-I). We predict 472 that it would take approximately three additional months for the oligodendrocyte population to be 473 fully regenerated, but this is only possible if a higher rate of oligodendrogenesis (~3.5%) than 474 observed in age-matched control brains (1.7%) is maintained. However, it is not clear whether 475


https://doi.org/10.1101/2020.03.09.983569


15

the endogenous pool of OPCs could maintain this higher rate of differentiation (and resulting 476 homeostatic OPC turnover) required for such a prolonged period of replacement. Although 477 recent studies indicate that GLI1-expressing glial progenitors positioned within germinal zones 478 along the lateral ventricles are mobilized in response to cuprizone-induced demyelination, 479 forming new OPCs that migrate and differentiate into oligodendrocytes with higher probability 480 than resident OPCs (Samanta et al., 2015), our results indicate that this recruitment is not 481 sufficient in the short term to overcome existing inhibitory barriers within cortical gray matter. 482 The inability to recover fully from a demyelinating event raises the possibility that 483 inflammation persists long after the initial trauma, or that OPCs in these regions are 484 permanently altered as a result of exposure to this environment, if only for a short time (Baxi et 485 al., 2015; Kirby et al., 2019). Like other progenitor cells, OPCs exhibit a decline in regenerative 486 potential with age and can undergo senescence, a process that may be accelerated by 487 exposure to inflammatory cytokines (Kirby et al., 2019; Neumann et al., 2019; Nicaise et al., 488 2019). It is also possible that there is a restricted time period during which OPCs can detect and 489 respond to myelin loss; if there are inherent limits on OPC mobilization, as suggested by the 490 uniform behavior of OPCs across cortical layers, then the inability to match the demand for new 491 cells early may lead to prolonged deficits. In this regard, evidence that certain aspects of the 492 inflammatory response strongly promote OPC differentiation (Kotter, Li, Zhao, & Franklin, 2006; 493 Miron et al., 2013; Ruckh et al., 2012) raises the possibility that there is a critical time window 494 for optimal repair. A more detailed spatial and cell-type specific profiling of inflammatory 495 changes in the cortex after oligodendrocyte death may help clarify the role of inflammatory 496 changes in this impaired regeneration. 497

An unexpected feature of oligodendrocyte regeneration in the cortex was that new cells 498 were not formed in the same locations as the prior oligodendrocytes, even though the high 499 density, regular spacing and dynamic motility of OPCs seem ideally suited to optimize 500 placement of new oligodendrocytes. Inhibitory factors generated as a consequence of 501 oligodendrocyte degeneration, such as myelin debris (Kotter et al., 2006), axonally expressed 502 factors (LINGO1) (Mi et al., 2005), extracellular matrix components (chondroitin sulfate 503 proteoglycans) (Lau et al., 2012), and reactive astrocytes (Back et al., 2005) may create a local 504 zone of exclusion at sites of cell death reducing the probability of OPC differentiation. 505

Complete reconstruction of individual oligodendrocytes revealed that regenerated cells 506 underwent a similar period of structural refinement over a period of ~ 10 days and ultimately 507 formed a comparable number of sheaths (Figure 3D,E, Supplementary Figure 4A-J). Although 508 oligodendrocytes have the capacity to form long processes, they did not always extend their 509


https://doi.org/10.1101/2020.03.09.983569


16

cytoplasmic processes to reach sites of original myelination, instead creating sheaths within a 510 local territory similar to that of normal oligodendrocytes (Figure 3D,I), with only a 15.5% 511 increase in total sheath length (Figure 3F). As regenerated oligodendrocytes are formed in an 512 environment with an apparent surplus of receptive axons, these findings suggest that the size 513 and shape of oligodendrocytes is profoundly limited by cell intrinsic mechanisms. 514 515 Reparative potential of surviving oligodendrocytes 516 Recent studies of postmortem human tissue from MS patients have raised the intriguing 517 possibility that remyelination may occur through the reformation of myelin sheaths by 518 oligodendrocytes that survive autoimmune attack, rather than from de novo oligodendrogenesis 519 (Yeung et al., 2019). This hypothesis is based on evidence that “shadow plaques”, which are 520 classically considered to represent partially remyelinated axons (Lassmann, Brück, Lucchinetti, 521 & Rodriguez, 1997), did not appear to contain many newly born oligodendrocytes, as assessed 522 using C14–based birth dating; the progressive decline in atmospheric C14 levels following the 523 cessation of atomic testing in the 1950s allow the date of last cell division to be estimated the 524 amount of C14 present. Although in vivo imaging studies indicate that oligodendrocytes can be 525 generated through direct differentiation of OPCs without cell division (Hughes et al., 2013), 526 potentially confounding measures of cell age based on proliferation, these results nevertheless 527 posit that significant new myelin can be created by existing oligodendrocytes. In our studies, the 528 few oligodendrocytes that survived cuprizone did not contribute substantial new myelin. 529 However, the cuprizone model does not fully recapitulate the pathology of cortical lesions 530 observed in MS. In particular, demyelination of the upper layers of the cortex in autopsy 531 samples from MS patients are correlated with regions of leptomeningeal inflammation, 532 composed of B and T cells that secrete cytotoxic cytokines and create a complex inflammatory 533 milieu (Howell et al., 2011; Roberta Magliozzi et al., 2007). Whether a cell-mediated immune 534 response, local release of cytotoxic compounds or environmental changes substantially shift the 535 burden of repair from OPCs to surviving oligodendrocytes in human MS remains to be explored. 536 537 Specificity of myelin repair 538 The regeneration of oligodendrocytes in different locations and the strong, cell intrinsic control of 539 cell size appear to constrain where myelin sheaths are formed. However, when a regenerated 540 oligodendrocyte had access to a territory that was previously myelinated, it was capable of 541 establishing sheaths along axons that were previously myelinated, indicating that the local 542 factors which initially influenced axon selection were retained after demyelination. In these 543


https://doi.org/10.1101/2020.03.09.983569


17

regions, myelin sheaths were often reformed at a similar position along axons, even when 544 sheaths were isolated from neighboring sheaths and therefore could have extended over a 545 much larger area. Although previous studies suggest that components of the NOR are 546 redistributed after demyelination in the PNS (England, Gamboni, Levinson, & Finger, 1990) and 547 in MS lesions (Coman et al., 2006; Craner et al., 2004; Dupree et al., 2004), our studies reveal 548

that βIV-spectrin, which forms a complex with Ankyrin-G to link voltage-gated sodium channels 549 to the actin cytoskeleton at NORs (Susuki et al., 2016), remains clustered (without flanking 550 myelin sheaths) with continued administration of cuprizone for up to six weeks (Figure 7D,F), 551 suggesting that axonal guideposts remain for many weeks after demyelination. As these studies 552 were performed using post-hoc immunostaining, we do not yet know if these βIV-spectrin 553 puncta are located at previous nodal positions; future longitudinal studies using fluorescently 554 tagged nodal components will help define the stability of NORs after demyelination. The time 555 course over which these components are removed from demyelinated axons could influence the 556 subsequent distribution of myelin, with prolonged demyelinating injuries leading to greater loss 557 of remyelination specificity. Stabilizing the structural elements of previously myelinated axonal 558 domains could therefore represent a potential target for remyelinating therapies. 559

Because territory overlap between original and regenerated oligodendrocytes was only 560 ~60%, considerable new axonal territory was accessed during the reparative period, leading to 561 novel sheaths that altered the global pattern of cortical myelin. However, even in areas where 562 the territories of regenerated and original oligodendrocytes overlapped, new myelin sheaths 563 were often formed on portions of axons that were not previously myelinated. These findings 564 highlight the probabilistic nature of myelination, which is influenced by many dynamic factors, 565 such as neural activity, axon size and metabolic state that could alter target selection and 566 stabilization of nascent sheaths (Klingseisen et al., 2019). 567 568 Implications for myelin repair and cognitive function 569 Oligodendrocytes perform crucial roles in the CNS, enhancing the propagation of action 570 potentials and reducing the metabolic cost to do so, providing metabolic support for axons far 571 removed from their cell bodies and controlling excitability by influencing the distribution of 572 voltage-gated channels and promoting clearance of extracellular potassium. Thus, the 573 reorganization of myelin that occurs in cortical circuits during regeneration may have profound 574 functional consequences on cognition and behavior. It will be important to determine whether 575 these functional aspects of oligodendrocyte-neuron interactions are restored following 576 remyelination. Our studies focused exclusively on regeneration in the somatosensory cortex of 577


https://doi.org/10.1101/2020.03.09.983569


18

young adult mice. It is not yet known if these changes mimic regeneration in other cortical 578 regions, or if the spatial and temporal aspects of myelin replacement in the cortex vary with age. 579 Moreover, due to limitations in resolving sheaths in deeper layers of the cortex, we do not yet 580 know whether regeneration deficits extend to layers IV-VI. It will also be critical to define the 581 spatial and temporal changes in myelin sheath thickness, as this varies considerably between 582 different axons and has been shown to be a substrate for plasticity in the cortex (Gibson et al., 583 2014). 584

Our analysis was restricted to discrete volumes in the cortical mantle. As we are not yet 585 able to monitor myelination patterns along the entire length of individual axons, it is possible that 586 the position of sheaths may change after regeneration, but that the overall myelin content along 587 a given axon is conserved. Alternatively, from a functional standpoint, it may not be necessary 588 to reform the precise pattern of myelination, as long as the relative amount of myelin along 589 different classes of neurons is preserved. At present, our ability to predict the consequences of 590 these changing myelin patterns and the spatial differences in oligodendrocyte regeneration are 591 limited by our knowledge about the function of myelin in cortical gray matter. As recent studies 592 suggest that even subtle changes in oligodendrogenesis can alter behavioral performance 593 (McKenzie et al., 2014; Xiao et al., 2016), the impact of these changes may be profound. The 594 ability to monitor the regeneration of myelin sheaths with high spatial and temporal resolution in 595 vivo within defined circuits provides new opportunities to evaluate the effectiveness of potential 596 therapeutic interventions (Deshmukh et al., 2013; Early et al., 2018; Mei et al., 2014; Rankin et 597 al., 2019) and a platform to explore the functional consequences of myelin reorganization. 598 599 600 MATERIALS AND METHODS 601 602 Animal care and use 603 Female and male adult mice were used for experiments and randomly assigned to experimental 604 groups. All mice were healthy and did not display any overt behavioral phenotypes, and no 605 animals were excluded from the analysis. Generation and genotyping of BAC transgenic lines 606 from Mobp-EGFP (Gensat) (Hughes et al., 2018) have been previously described. Mice were 607 maintained on a 12-h light/dark cycle, housed in groups no larger than 5, and food and water 608 were provided ad libitum (except during cuprizone-administration, see below). All animal 609 experiments were performed in strict accordance with protocols approved by the Animal Care 610 and Use Committee at Johns Hopkins University. 611


https://doi.org/10.1101/2020.03.09.983569


19

612 Cranial Windows 613 Cranial windows were prepared as previously described (Holtmaat et al., 2012; Hughes et al., 614 2018). Briefly, mice 7 to 10 weeks old were anesthetized with isoflurane (induction, 5%; 615 maintenance, 1.5-2%, mixed with 0.5L/min O2), and their body temperature was maintained at 616 37o C with a thermostat-controlled heating plate. The skin over the right hemisphere was 617 removed and the skull cleaned. A 2 x 2 or 3 x 3 mm region of skull over somatosensory cortex (-618 1.5 mm posterior and 3.5 mm lateral from bregma) was removed using a high-speed dental drill. 619 A piece of cover glass (VWR, No. 1) was placed in the craniotomy and sealed with VetBond 620 (3M), then dental cement (C&B Metabond) and a custom metal plate with a central hole was 621 attached to the skull for head stabilization. 622 623 In vivo two photon microscopy 624 In vivo imaging sessions began 2 to 3 weeks after cranial window procedure (Baseline). After 625 the baseline imaging session, mice were randomly assigned to cuprizone or control conditions. 626 During imaging sessions, mice were anesthetized with isoflurane and immobilized by attaching 627 the head plate to a custom stage. Images were collected using a Zeiss LSM 710 microscope 628 equipped with a GaAsP detector using a mode-locked Ti:sapphire laser (Coherent Ultra) tuned 629 to 920 nm. The average power at the sample during imaging was < 30 mW. Vascular landmarks 630 were used to identify the same cortical area over longitudinal imaging sessions. Image stacks 631 were 425 µm x 425 µm x 110 µm (2048 x 2048 pixels, corresponding to cortical layer I, Zeiss 632

20x objective) , 425 µm x 425 µm x 550 µm (1024 x 1024 pixels) or 850 µm x 850 µm x 550 µm 633 (1024 x 1024 pixels; corresponding to layers I – IV), relative the cortical surface. Mice were 634 imaged every 1 to 7 days, for up to 15 weeks. 635 636 Cuprizone administration 637 At 9 to 11 weeks of age, male and female Mobp-EGFP mice were fed a diet of milled, irradiated 638 18% protein rodent diet (Teklad Global) alone (control) or supplemented with 0.2% w/w 639 bis(cyclohexanone) oxaldihydrazone (Cuprizone, Sigma-Aldrich) in custom gravity-fed food 640 dispensers for 3 to 6 weeks. Both control and experimental condition mice were returned to 641 regular pellet diet during the recovery period (Baxi et al., 2017). 642 643 Immunohistochemistry 644


https://doi.org/10.1101/2020.03.09.983569


20

Mice were deeply anesthetized with sodium pentobarbital (100 mg/kg b.w.) and perfused 645 transcardially with 4% paraformaldehyde (PFA in 0.1 M phosphate buffer, pH 7.4). Brains were 646 then post-fixed in 4% PFA for 12 to 18 hours, depending on antibody sensitivity to fixation, 647 before being transferred to a 30% sucrose solution (in PBS, pH 7.4). For horizontal sections, 648 cortices were flat-mounted between glass slides and postfixed in 4% PFA for 6 to 12 hours at 4o 649 C, transferred to 30% sucrose solution (in PBS, pH 7.4). Tissue was stored at 4o C for more 650 than 48 h before sectioning. Brains were extracted, frozen in TissueTek, sectioned (-1.5 mm 651 posterior and 3.5 mm lateral from bregma) at 30 to 50 µm thickness on a cryostat (Thermo 652 Scientific Microm HM 550) at -20o C. Immunohistochemistry was performed on free-floating 653 sections. Sections were preincubated in blocking solution (5% normal donkey serum, 0.3% 654 Triton X-100 in PBS, pH 7.4) for 1 or 2 hours at room temperature, then incubated for 24 to 48 655 hours at 4o C or room temperature in primary antibody (listed in Key Resources Table). 656 Secondary antibody (see Key Resources Table) incubation was performed at room temperature 657 for 2 to 4 hours. Sections were mounted on slides with Aqua Polymount (Polysciences). Images 658 were acquired using either an epifluorescence microscope (Zeiss Axio-imager M1) with 659 Axiovision software (Zeiss) or a confocal laser-scanning microscope (Zeiss LSM 510 Meta; 660 Zeiss LSM 710; Zeiss LSM 880). For glial cell counts, individual images of coronal sections 661 were quantified by a blinded observer for number of NG2+, GFAP+ and GFP+ cells within a 425 662 μm x 500 µm region, and divided into 425 μm x 100 µm zones from the pial surface 663 (Supplementary Figure 1). Immunostaining for nodal components was performed as above, 664 except mice were transcardially perfused with PBS only and post-fixed in 4% PFA for 50 665 minutes. 666 667 Image processing and analysis 668 Image stacks and time series were analyzed using FIJI/ImageJ. For presentation in figures, 669 image brightness and contrast levels were adjusted for clarity. Myelin sheath images were 670 additionally de-noised with a 3-D median filter (radius 0.5 to 1.5 pixels). Longitudinal image 671 stacks were registered using FIJI plugin “Correct 3D Drift” (Parslow, Cardona, & Bryson-672 Richardson, 2014) and then randomized for analysis by a blinded observer. 673

674 Cell tracking 675 Individual oligodendrocytes were followed in four dimensions using custom FIJI scripts (Hughes 676 et al., 2018) or with SyGlass (IstoVisio) virtual reality software by defining individual EGFP+ cell 677 bodies at each time point, recording xyz coordinates, and defining cellular behavior (new, lost, 678


https://doi.org/10.1101/2020.03.09.983569


21

or stable cells). Oligodendrocytes that were characterized as “lost” initially lost EGFP signal in 679 processes and myelin sheaths, before complete loss of signal from the cell body position. A 680 “new” oligodendrocyte appeared with novel processes and internodes absent in baseline 681 images. Dynamics of cell body positions were analyzed with custom MATLAB scripts, and 682 cross-time point comparisons of 3-D coordinates were corrected by adding the average vector 683 of movement of all cells between those timepoints (to account for imperfect image registration 684 and expansion/contraction of brain volume over time). For quantification between different 100 685 µm depths, cells were binned between planes horizontal to the plane of the pia, and included 686 cells were found by Delaunay triangulation. 687

688 Myelin sheath analysis 689 Registered longitudinal in vivo Z-stacks collected from Mobp-EGFP mice were acquired using 690 two-photon microscopy. Similar to that described previously (Hughes et al., 2018), all myelin 691 sheaths within a volume of 100 µm x 100 µm x 100 µm from the pial surface were traced in FIJI 692 using Simple Neurite Tracer (Longair, Baker, & Armstrong, 2011) at the baseline or final 693 recovery time-point. Then, using registered time-series images from baseline to final recovery 694 time-point, each myelin sheath was categorized as having 0, 1 or 2 myelin sheath neighbors 695 (Figure 4), whether it was stable (connected via cytoplasmic process to same cell at baseline 696 and at 5 weeks recovery), lost (present at baseline, but not at 5 weeks recovery), replaced (≥ 697 50% of the original sheath length was replaced by a sheath connected via cytoplasmic process 698 to a regenerated oligodendrocyte), or novel (a sheath not present at baseline that was 699 connected to a regenerated oligodendrocyte). If it was a stable or replaced myelin sheath, we 700 determined whether the baseline myelin sheath had the same or different number of 701 neighboring myelin sheaths. Myelin sheaths within the field that could not be definitively 702 categorized were classified as “undefined”. Myelin paranodes were identified by increased 703 EGFP fluorescence intensity (Hughes et al., 2018). Nodes of Ranvier were confirmed by plotting 704 an intensity profile across the putative node; if the profile consisted of two local maxima 705 separated by a minimum less than that of the internode, and the length of the gap between 706 EGFP+ processes was < 5 µm, the structure was considered a node. For each field, myelin 707 sheaths were traced by one investigator and independently assessed by a second investigator. 708 709 Analysis of temporal and spatial dynamics of individual oligodendrocytes 710 Registered longitudinal in vivo Z-stacks collected from Mobp-EGFP mice were acquired using 711 two-photon microscopy every 1-3 days to follow the dynamics of newly formed mature 712


https://doi.org/10.1101/2020.03.09.983569


22

oligodendrocytes within cortical layer I at high resolution (200 - 400 µm x-y, 100 - 120 µm z; 713 2048 x 2048 pixels). Using images from day of appearance, all processes originating from the 714 cell body, branch points, and individual myelin sheaths were traced in FIJI using Simple Neurite 715 Tracer (Longair et al., 2011). Traced segments were put through a smoothing function prior to 716 length calculations to reduce artifacts of jagged traces. The fate of each process and myelin 717 sheath (stable, lost) and changes in length (stable, growth, retraction) were determined for up to 718 14 days per cell. 719 720 Single oligodendrocyte territory analysis 721 3-D coordinates of traces from all sheaths of a single oligodendrocyte were averaged to find the 722 center of mass (to reduce artifactual territory volume that would be above the pia resulting from 723 centering at the cell body). From this center, ellipsoid volumes were calculated from x-y and z 724 radii, working backwards from the distance of the furthest sheath voxel in 1-µm increments in 725 each dimension. The ellipsoid dimensions for each traced cell was determined by the 726 combination of x-y and z radii that produced the smallest volume containing ≥ 80% of the sheath 727

voxels for that cell. Voxels within the ellipsoid were calculated by 𝑥𝑥𝑖𝑖𝑟𝑟𝑥𝑥

+ 𝑦𝑦𝑖𝑖𝑟𝑟𝑦𝑦

+ 𝑧𝑧𝑖𝑖𝑟𝑟𝑧𝑧≤ 1 where (𝑥𝑥𝑖𝑖 ,𝑦𝑦𝑖𝑖 , 𝑧𝑧𝑖𝑖) 728

is a voxel belonging to a sheath, and 𝑟𝑟𝑥𝑥, 𝑟𝑟𝑦𝑦, and 𝑟𝑟𝑧𝑧 are the radii being tested. Because all traced 729

cell morphologies were not found to be significantly different from radial, the x and y radii were 730 held equivalent. The average x-y and z radii across all traced baseline, control, or regenerated 731 cells were used as standard dimensions for an oligodendrocyte “territory” in subsequent 732 calculations. 733 734 Total territory volume for a single time-point (either baseline or 5 weeks recovery) was 735 calculated in the following manner. In a 3-D matrix with dimensions, 425 μm x 425 μm x 33 μm 736 corresponding to the number of voxels in the top zone of the imaged volume, all voxels 737 contained within the territories of each cell were represented as 1’s, and all voxels outside of 738 cell territories were represented as 0’s. Voxels within territories of more than one cell had values 739 equal to the number of territories they were within. Total baseline volume “replaced” by 740 regenerated territories was calculated as: ((5-week recovery matrix – baseline matrix) ≤ 0) + 741 baseline matrix. The total proportion of baseline volume overlapped by regenerated territories 742 was then calculated as: “replaced volume” / baseline volume. Because each imaged region had 743 varying numbers of oligodendrocytes at baseline and regenerated oligodendrocytes in recovery, 744


https://doi.org/10.1101/2020.03.09.983569


23

final territory overlap proportions were scaled in the following manner: total overlap × (number 745

of cells @ baseline ÷ number of cells @ 5 weeks recovery). 746 747 Analysis of nodal components 748 Following immunostaining, images were acquired at 63x on a Zeiss 880 microscope at high 749 resolution (135 x 135 x 40 µm, 2048 x 2048 pixels). Images were processed in FIJI (background 750 subtraction, rolling ball 60 pixels; 3-D median filter, 1.2 pixels XY 0.5 pixels Z). Images were 751 acquired from sections immunostained for Caspr and βIV-spectrin, or both βIV-spectrin and 752 Ankyrin-G. For images with both nodal components labeled, we found complete overlap of 753 signal. In these cases, images were processed with “Image Calculator” in FIJI to reduce noise 754 between channels, subtracting Ankyrin-G from βIV-spectrin signal. Images were imported into 755 IMARIS and 3-D positions of all nodal signal and paranodal puncta (Caspr) were resolved with 756 low- and high-pass filters and “Surface” and “Spot” functions. Axon initial segments were 757 excluded using a size cutoff of ≤ 6 µm. Custom MATLAB scripts were used to detect proximity 758 (nearest neighbor Euclidian distance, threshold 3.5 µm radius) of nodal puncta to Caspr puncta 759

to determine proportions of nodes of Ranvier (2 Caspr puncta within 3.5 µm), heminodes (1 760

Caspr punctum within 3.5 µm), or isolated (no nearby Caspr puncta). Coordinates from one 761 channel were rotated 90 degrees in the x-y plane before running the proximity analysis again to 762 confirm observed proportions were not due to chance. 763 764 Statistical analysis 765 No statistical tests were used to predetermine sample sizes, but our sample sizes are similar to 766 those reported in previous publications (Hughes et al., 2018). Statistical analyses were 767 performed with MATLAB (Mathworks) or Excel (Microsoft). Significance was typically 768 determined using N-way ANOVA test with Bonferroni correction for multiple comparisons. Each 769 figure legend otherwise contains the statistical tests used to measure significance and the 770 corresponding significance level (p value). N represents the number of animals used in each 771 experiment. Data are reported as mean ± SEM and p < 0.05 was considered statistically 772 significant. 773 774 Data availability 775 All published code, tools, and reagents will be shared on an unrestricted basis; requests should 776 be directed to the corresponding authors. 777 778


https://doi.org/10.1101/2020.03.09.983569


24

REFERENCES 779 Araújo, S. E. S., Mendonça, H. R., Wheeler, N. A., Campello-Costa, P., Jacobs, K. M., Gomes, 780

F. C. A., … Fuss, B. (2017). Inflammatory demyelination alters subcortical visual circuits. 781 Journal of Neuroinflammation, 14(1), 162. https://doi.org/10.1186/s12974-017-0936-0 782

Auer, F., Vagionitis, S., & Czopka, T. (2018). Evidence for Myelin Sheath Remodeling in the 783 CNS Revealed by In Vivo Imaging. Current Biology : CB, 28(4), 549-559.e3. 784 https://doi.org/10.1016/j.cub.2018.01.017 785

Back, S. A., Tuohy, T. M. F., Chen, H., Wallingford, N., Craig, A., Struve, J., … Sherman, L. S. 786 (2005). Hyaluronan accumulates in demyelinated lesions and inhibits oligodendrocyte 787 progenitor maturation. Nature Medicine, 11(9), 966–972. https://doi.org/10.1038/nm1279 788

Baxi, E. G., DeBruin, J., Jin, J., Strasburger, H. J., Smith, M. D., Orthmann-Murphy, J. L., … 789 Calabresi, P. A. (2017). Lineage tracing reveals dynamic changes in oligodendrocyte 790 precursor cells following cuprizone-induced demyelination. GLIA, 65(12), 2087–2098. 791 https://doi.org/10.1002/glia.23229 792

Baxi, E. G., DeBruin, J., Tosi, D. M., Grishkan, I. V., Smith, M. D., Kirby, L. A., … Gocke, A. R. 793 (2015). Transfer of Myelin-Reactive Th17 Cells Impairs Endogenous Remyelination in the 794 Central Nervous System of Cuprizone-Fed Mice. Journal of Neuroscience, 35(22), 8626–795 8639. https://doi.org/10.1523/JNEUROSCI.3817-14.2015 796

Beck, E. S., Sati, P., Sethi, V., Kober, T., Dewey, B., Bhargava, P., … Reich, D. S. (2018). 797 Improved Visualization of Cortical Lesions in Multiple Sclerosis Using 7T MP2RAGE. 798 American Journal of Neuroradiology, 39(3), 459–466. https://doi.org/10.3174/ajnr.A5534 799

Biteau, B., Hochmuth, C. E., & Jasper, H. (2011). Maintaining Tissue Homeostasis: Dynamic 800 Control of Somatic Stem Cell Activity. Cell Stem Cell, 9(5), 402–411. 801 https://doi.org/10.1016/j.stem.2011.10.004 802

Bock, D. D., Lee, W.-C. A., Kerlin, A. M., Andermann, M. L., Hood, G., Wetzel, A. W., … Reid, 803 R. C. (2011). Network anatomy and in vivo physiology of visual cortical neurons. Nature, 804 471(7337), 177–182. https://doi.org/10.1038/nature09802 805

Calabrese, M., Poretto, V., Favaretto, A., Alessio, S., Bernardi, V., Romualdi, C., … Gallo, P. 806 (2012). Cortical lesion load associates with progression of disability in multiple sclerosis. 807 Brain, 135(10), 2952–2961. https://doi.org/10.1093/brain/aws246 808

Chang, A., Nishiyama, A., Peterson, J., Prineas, J., & Trapp, B. D. (2000). NG2-positive 809 oligodendrocyte progenitor cells in adult human brain and multiple sclerosis lesions. The 810 Journal of Neuroscience : The Official Journal of the Society for Neuroscience, 20(17), 811 6404–6412. https://doi.org/10.1523/JNEUROSCI.20-17-06404.2000 812


https://doi.org/10.1101/2020.03.09.983569


25

Chang, A., Staugaitis, S. M., Dutta, R., Batt, C. E., Easley, K. E., Chomyk, A. M., … Trapp, B. D. 813 (2012). Cortical remyelination: A new target for repair therapies in multiple sclerosis. 814 Annals of Neurology, 72(6), 918–926. https://doi.org/10.1002/ana.23693 815

Chong, S. Y. C., Rosenberg, S. S., Fancy, S. P. J., Zhao, C., Shen, Y.-A. A., Hahn, A. T., … 816 Chan, J. R. (2012). Neurite outgrowth inhibitor Nogo-A establishes spatial segregation and 817 extent of oligodendrocyte myelination. Proceedings of the National Academy of Sciences of 818 the United States of America, 109(4), 1299–1304. 819 https://doi.org/10.1073/pnas.1113540109 820

Coman, I., Aigrot, M. S., Seilhean, D., Reynolds, R., Girault, J. A., Zalc, B., & Lubetzki, C. 821 (2006). Nodal, paranodal and juxtaparanodal axonal proteins during demyelination and 822 remyelination in multiple sclerosis. Brain : A Journal of Neurology, 129(Pt 12), 3186–3195. 823 https://doi.org/10.1093/brain/awl144 824

Craner, M. J., Newcombe, J., Black, J. A., Hartle, C., Cuzner, M. L., & Waxman, S. G. (2004). 825 Molecular changes in neurons in multiple sclerosis: Altered axonal expression of Nav1.2 826 and Nav1.6 sodium channels and Na+/Ca2+ exchanger. Proceedings of the National 827 Academy of Sciences, 101(21), 8168–8173. https://doi.org/10.1073/pnas.0402765101 828

Czopka, T., Ffrench-Constant, C., & Lyons, D. A. (2013). Individual oligodendrocytes have only 829 a few hours in which to generate new myelin sheaths in vivo. Developmental Cell, 25(6), 830 599–609. https://doi.org/10.1016/j.devcel.2013.05.013 831

Deshmukh, V. A., Tardif, V., Lyssiotis, C. A., Green, C. C., Kerman, B., Kim, H. J., … Lairson, L. 832 L. (2013). A regenerative approach to the treatment of multiple sclerosis. Nature, 833 502(7471), 327–332. https://doi.org/10.1038/nature12647 834

Dimou, L., Simon, C., Kirchhoff, F., Takebayashi, H., & Götz, M. (2008). Progeny of Olig2-835 expressing progenitors in the gray and white matter of the adult mouse cerebral cortex. 836 The Journal of Neuroscience : The Official Journal of the Society for Neuroscience, 28(41), 837 10434–10442. https://doi.org/10.1523/JNEUROSCI.2831-08.2008 838

Dupree, J. L., Mason, J. L., Marcus, J. R., Stull, M., Levinson, R., Matsushima, G. K., & Popko, 839 B. (2004). Oligodendrocytes assist in the maintenance of sodium channel clusters 840 independent of the myelin sheath. Neuron Glia Biology, 1(3), 179–192. 841 https://doi.org/10.1017/S1740925X04000304 842

Dutta, D. J., Woo, D. H., Lee, P. R., Pajevic, S., Bukalo, O., Huffman, W. C., … Fields, R. D. 843 (2018). Regulation of myelin structure and conduction velocity by perinodal astrocytes. 844 Proceedings of the National Academy of Sciences of the United States of America, 845 115(46), 11832–11837. https://doi.org/10.1073/pnas.1811013115 846


https://doi.org/10.1101/2020.03.09.983569


26

Early, J. J., Cole, K. L., Williamson, J. M., Swire, M., Kamadurai, H., Muskavitch, M., & Lyons, 847 D. A. (2018). An automated high-resolution in vivo screen in zebrafish to identify chemical 848 regulators of myelination. ELife, 7. https://doi.org/10.7554/eLife.35136 849

England, J. D., Gamboni, F., Levinson, S. R., & Finger, T. E. (1990). Changed distribution of 850 sodium channels along demyelinated axons. Proceedings of the National Academy of 851 Sciences of the United States of America, 87(17), 6777–6780. Retrieved from 852 http://www.ncbi.nlm.nih.gov/pubmed/2168559 853

Filippi, M., Evangelou, N., Kangarlu, A., Inglese, M., Mainero, C., Horsfield, M. A., & Rocca, M. 854 A. (2014). Ultra-high-field MR imaging in multiple sclerosis. Journal of Neurology, 855 Neurosurgery, and Psychiatry, 85(1), 60–66. https://doi.org/10.1136/jnnp-2013-305246 856

Gibson, E. M., Purger, D., Mount, C. W., Goldstein, A. K., Lin, G. L., Wood, L. S., … Monje, M. 857 (2014). Neuronal activity promotes oligodendrogenesis and adaptive myelination in the 858 mammalian brain. Science (New York, N.Y.), 344(6183). 859 https://doi.org/10.1126/science.1252304 860

Gouw, A. A., Seewann, A., Vrenken, H., van der Flier, W. M., Rozemuller, J. M., Barkhof, F., … 861 Geurts, J. J. G. (2008). Heterogeneity of white matter hyperintensities in Alzheimer’s 862 disease: post-mortem quantitative MRI and neuropathology. Brain, 131(12), 3286–3298. 863 https://doi.org/10.1093/brain/awn265 864

Gudi, V., Moharregh-Khiabani, D., Skripuletz, T., Koutsoudaki, P. N., Kotsiari, A., Skuljec, J., … 865 Stangel, M. (2009). Regional differences between grey and white matter in cuprizone 866 induced demyelination. Brain Research, 1283, 127–138. 867 https://doi.org/10.1016/j.brainres.2009.06.005 868

Herranz, E., Louapre, C., Treaba, C. A., Govindarajan, S. T., Ouellette, R., Mangeat, G., … 869 Mainero, C. (2019). Profiles of cortical inflammation in multiple sclerosis by 11C-PBR28 870 MR-PET and 7 Tesla imaging. Multiple Sclerosis (Houndmills, Basingstoke, England), 871 1352458519867320. https://doi.org/10.1177/1352458519867320 872

Hill, R. A., Li, A. M., & Grutzendler, J. (2018). Lifelong cortical myelin plasticity and age-related 873 degeneration in the live mammalian brain. Nature Neuroscience, 21(5), 683–695. 874 https://doi.org/10.1038/s41593-018-0120-6 875

Holtmaat, A., de Paola, V., Wilbrecht, L., Trachtenberg, J. T., Svoboda, K., & Portera-Cailliau, 876 C. (2012). Imaging neocortical neurons through a chronic cranial window. Cold Spring 877 Harbor Protocols, 2012(6), 694–701. https://doi.org/10.1101/pdb.prot069617 878

Howell, O. W., Reeves, C. A., Nicholas, R., Carassiti, D., Radotra, B., Gentleman, S. M., … 879 Reynolds, R. (2011). Meningeal inflammation is widespread and linked to cortical 880


https://doi.org/10.1101/2020.03.09.983569


27

pathology in multiple sclerosis. Brain, 134(9), 2755–2771. 881 https://doi.org/10.1093/brain/awr182 882

Hughes, E. G., Kang, S. H., Fukaya, M., & Bergles, D. E. (2013). Oligodendrocyte progenitors 883 balance growth with self-repulsion to achieve homeostasis in the adult brain. Nature 884 Neuroscience, 16(6), 668–676. https://doi.org/10.1038/nn.3390 885

Hughes, E. G., Orthmann-Murphy, J. L., Langseth, A. J., & Bergles, D. E. (2018). Myelin 886 remodeling through experience-dependent oligodendrogenesis in the adult somatosensory 887 cortex. Nature Neuroscience, 21(5), 696–706. https://doi.org/10.1038/s41593-018-0121-5 888

Ihara, M., Polvikoski, T. M., Hall, R., Slade, J. Y., Perry, R. H., Oakley, A. E., … Kalaria, R. N. 889 (2010). Quantification of myelin loss in frontal lobe white matter in vascular dementia, 890 Alzheimer’s disease, and dementia with Lewy bodies. Acta Neuropathologica, 119(5), 579–891 589. https://doi.org/10.1007/s00401-009-0635-8 892

Jeffery, N. D., & Blakemore, W. F. (1995). Remyelination of mouse spinal cord axons 893 demyelinated by local injection of lysolecithin. Journal of Neurocytology, 24(10), 775–781. 894 https://doi.org/10.1007/bf01191213 895

Kang, S. H., Fukaya, M., Yang, J. K., Rothstein, J. D., & Bergles, D. E. (2010). NG2+ CNS glial 896 progenitors remain committed to the oligodendrocyte lineage in postnatal life and following 897 neurodegeneration. Neuron, 68(4), 668–681. 898

Kidd, D., Barkhof, F., McConnell, R., Algra, P. R., Allen, I. V, & Revesz, T. (1999). Cortical 899 lesions in multiple sclerosis. Brain : A Journal of Neurology, 122 ( Pt 1), 17–26. 900 https://doi.org/10.1093/brain/122.1.17 901

Kilsdonk, I. D., de Graaf, W. L., Soriano, A. L., Zwanenburg, J. J., Visser, F., Kuijer, J. P. A., … 902 Wattjes, M. P. (2013). Multicontrast MR imaging at 7T in multiple sclerosis: highest lesion 903 detection in cortical gray matter with 3D-FLAIR. AJNR. American Journal of 904 Neuroradiology, 34(4), 791–796. https://doi.org/10.3174/ajnr.A3289 905

Kirby, L., Jin, J., Cardona, J. G., Smith, M. D., Martin, K. A., Wang, J., … Calabresi, P. A. 906 (2019). Oligodendrocyte precursor cells present antigen and are cytotoxic targets in 907 inflammatory demyelination. Nature Communications, 10(1), 3887. 908 https://doi.org/10.1038/s41467-019-11638-3 909

Klingseisen, A., Ristoiu, A.-M., Kegel, L., Sherman, D. L., Rubio-Brotons, M., Almeida, R. G., … 910 Lyons, D. A. (2019). Oligodendrocyte Neurofascin Independently Regulates Both Myelin 911 Targeting and Sheath Growth in the CNS. Developmental Cell, 51(6), 730-744.e6. 912 https://doi.org/10.1016/j.devcel.2019.10.016 913

Kotter, M. R., Li, W.-W., Zhao, C., & Franklin, R. J. M. (2006). Myelin Impairs CNS 914


https://doi.org/10.1101/2020.03.09.983569


28

Remyelination by Inhibiting Oligodendrocyte Precursor Cell Differentiation. Journal of 915 Neuroscience, 26(1), 328–332. https://doi.org/10.1523/JNEUROSCI.2615-05.2006 916

Kutzelnigg, A. (2005). Cortical demyelination and diffuse white matter injury in multiple sclerosis. 917 Brain, 128(11), 2705–2712. https://doi.org/10.1093/brain/awh641 918

Lassmann, H., Brück, W., Lucchinetti, C., & Rodriguez, M. (1997). Remyelination in multiple 919 sclerosis. Multiple Sclerosis Journal, 3(2), 133–136. 920 https://doi.org/10.1177/135245859700300213 921

Lau, L. W., Keough, M. B., Haylock-Jacobs, S., Cua, R., Döring, A., Sloka, S., … Yong, V. W. 922 (2012). Chondroitin sulfate proteoglycans in demyelinated lesions impair remyelination. 923 Annals of Neurology, 72(3), 419–432. https://doi.org/10.1002/ana.23599 924

Lodato, S., & Arlotta, P. (2015). Generating neuronal diversity in the mammalian cerebral 925 cortex. Annual Review of Cell and Developmental Biology, 31, 699–720. 926 https://doi.org/10.1146/annurev-cellbio-100814-125353 927

Longair, M. H., Baker, D. A., & Armstrong, J. D. (2011). Simple Neurite Tracer: open source 928 software for reconstruction, visualization and analysis of neuronal processes. 929 Bioinformatics (Oxford, England), 27(17), 2453–2454. 930 https://doi.org/10.1093/bioinformatics/btr390 931

Lucchinetti, C. F., Popescu, B. F. G., Bunyan, R. F., Moll, N. M., Roemer, S. F., Lassmann, H., 932 … Ransohoff, R. M. (2011). Inflammatory cortical demyelination in early multiple sclerosis. 933 The New England Journal of Medicine, 365(23), 2188–2197. 934 https://doi.org/10.1056/NEJMoa1100648 935

Magliozzi, R., Reynolds, R., & Calabrese, M. (2018). MRI of cortical lesions and its use in 936 studying their role in MS pathogenesis and disease course. Brain Pathology, 28(5), 735–937 742. https://doi.org/10.1111/bpa.12642 938

Magliozzi, Roberta, Howell, O., Vora, A., Serafini, B., Nicholas, R., Puopolo, M., … Aloisi, F. 939 (2007). Meningeal B-cell follicles in secondary progressive multiple sclerosis associate with 940 early onset of disease and severe cortical pathology. Brain : A Journal of Neurology, 130(Pt 941 4), 1089–1104. https://doi.org/10.1093/brain/awm038 942

Matsushima, G. K., & Morell, P. (2001). The neurotoxicant, cuprizone, as a model to study 943 demyelination and remyelination in the central nervous system. Brain Pathology (Zurich, 944 Switzerland), 11(1), 107–116. https://doi.org/10.1111/j.1750-3639.2001.tb00385.x 945

McKenzie, I. A., Ohayon, D., Li, H., de Faria, J. P., Emery, B., Tohyama, K., & Richardson, W. 946 D. (2014). Motor skill learning requires active central myelination. Science (New York, 947 N.Y.), 346(6207), 318–322. https://doi.org/10.1126/science.1254960 948


https://doi.org/10.1101/2020.03.09.983569


29

Mei, F., Fancy, S. P. J., Shen, Y.-A. A., Niu, J., Zhao, C., Presley, B., … Chan, J. R. (2014). 949 Micropillar arrays as a high-throughput screening platform for therapeutics in multiple 950 sclerosis. Nature Medicine, 20(8), 954–960. https://doi.org/10.1038/nm.3618 951

Mi, S., Miller, R. H., Lee, X., Scott, M. L., Shulag-Morskaya, S., Shao, Z., … Pepinsky, R. B. 952 (2005). LINGO-1 negatively regulates myelination by oligodendrocytes. Nature 953 Neuroscience, 8(6), 745–751. https://doi.org/10.1038/nn1460 954

Micheva, K. D., Wolman, D., Mensh, B. D., Pax, E., Buchanan, J., Smith, S. J., & Bock, D. D. 955 (2016). A large fraction of neocortical myelin ensheathes axons of local inhibitory neurons. 956 ELife, 5. https://doi.org/10.7554/eLife.15784 957

Miron, V. E., Boyd, A., Zhao, J.-W., Yuen, T. J., Ruckh, J. M., Shadrach, J. L., … Ffrench-958 Constant, C. (2013). M2 microglia and macrophages drive oligodendrocyte differentiation 959 during CNS remyelination. Nature Neuroscience, 16(9), 1211–1218. 960 https://doi.org/10.1038/nn.3469 961

Murtie, J. C., Macklin, W. B., & Corfas, G. (2007). Morphometric analysis of oligodendrocytes in 962 the adult mouse frontal cortex. Journal of Neuroscience Research, 85(10), 2080–2086. 963 https://doi.org/10.1002/jnr.21339 964

Neumann, B., Baror, R., Zhao, C., Segel, M., Dietmann, S., Rawji, K. S., … Franklin, R. J. M. 965 (2019). Metformin Restores CNS Remyelination Capacity by Rejuvenating Aged Stem 966 Cells. Cell Stem Cell, 25(4), 473-485.e8. https://doi.org/10.1016/j.stem.2019.08.015 967

Nicaise, A. M., Wagstaff, L. J., Willis, C. M., Paisie, C., Chandok, H., Robson, P., … Crocker, S. 968 J. (2019). Cellular senescence in progenitor cells contributes to diminished remyelination 969 potential in progressive multiple sclerosis. Proceedings of the National Academy of 970 Sciences, 116(18), 9030–9039. https://doi.org/10.1073/pnas.1818348116 971

Nielsen, A. S., Kinkel, R. P., Madigan, N., Tinelli, E., Benner, T., & Mainero, C. (2013). 972 Contribution of cortical lesion subtypes at 7T MRI to physical and cognitive performance in 973 MS. Neurology, 81(7), 641–649. https://doi.org/10.1212/WNL.0b013e3182a08ce8 974

Oh, J., Ontaneda, D., Azevedo, C., Klawiter, E. C., Absinta, M., Arnold, D. L., … North American 975 Imaging in Multiple Sclerosis Cooperative. (2019). Imaging outcome measures of 976 neuroprotection and repair in MS. Neurology, 92(11), 10.1212/WNL.0000000000007099. 977 https://doi.org/10.1212/WNL.0000000000007099 978

Parslow, A., Cardona, A., & Bryson-Richardson, R. J. (2014). Sample drift correction following 979 4D confocal time-lapse imaging. Journal of Visualized Experiments : JoVE, (86). 980 https://doi.org/10.3791/51086 981

Peterson, J. W., Bö, L., Mörk, S., Chang, A., & Trapp, B. D. (2001). Transected neurites, 982


https://doi.org/10.1101/2020.03.09.983569


30

apoptotic neurons, and reduced inflammation in cortical multiple sclerosis lesions. Annals 983 of Neurology, 50(3), 389–400. Retrieved from 984 http://www.ncbi.nlm.nih.gov/pubmed/11558796 985

Rankin, K. A., Mei, F., Kim, K., Shen, Y.-A. A., Mayoral, S. R., Desponts, C., … Bove, R. (2019). 986 Selective Estrogen Receptor Modulators Enhance CNS Remyelination Independent of 987 Estrogen Receptors. The Journal of Neuroscience : The Official Journal of the Society for 988 Neuroscience, 39(12), 2184–2194. https://doi.org/10.1523/JNEUROSCI.1530-18.2019 989

Ruckh, J. M., Zhao, J.-W., Shadrach, J. L., van Wijngaarden, P., Rao, T. N., Wagers, A. J., & 990 Franklin, R. J. M. (2012). Rejuvenation of regeneration in the aging central nervous 991 system. Cell Stem Cell, 10(1), 96–103. https://doi.org/10.1016/j.stem.2011.11.019 992

Samanta, J., Grund, E. M., Silva, H. M., Lafaille, J. J., Fishell, G., & Salzer, J. L. (2015). 993 Inhibition of Gli1 mobilizes endogenous neural stem cells for remyelination. Nature, 994 526(7573), 448–452. https://doi.org/10.1038/nature14957 995

Skripuletz, T., Lindner, M., Kotsiari, A., Garde, N., Fokuhl, J., Linsmeier, F., … Stangel, M. 996 (2008). Cortical demyelination is prominent in the murine cuprizone model and is strain-997 dependent. The American Journal of Pathology, 172(4), 1053–1061. 998 https://doi.org/10.2353/ajpath.2008.070850 999

Stedehouder, J., Couey, J. J., Brizee, D., Hosseini, B., Slotman, J. A., Dirven, C. M. F., … 1000 Kushner, S. A. (2017). Fast-spiking Parvalbumin Interneurons are Frequently Myelinated in 1001 the Cerebral Cortex of Mice and Humans. Cerebral Cortex (New York, N.Y. : 1991), 27(10), 1002 5001–5013. https://doi.org/10.1093/cercor/bhx203 1003

Stedehouder, J., & Kushner, S. A. (2017). Myelination of parvalbumin interneurons: a 1004 parsimonious locus of pathophysiological convergence in schizophrenia. Molecular 1005 Psychiatry, 22(1), 4–12. https://doi.org/10.1038/mp.2016.147 1006

Su, Z., Yuan, Y., Chen, J., Zhu, Y., Qiu, Y., Zhu, F., … He, C. (2011). Reactive astrocytes inhibit 1007 the survival and differentiation of oligodendrocyte precursor cells by secreted TNF-α. 1008 Journal of Neurotrauma, 28(6), 1089–1100. https://doi.org/10.1089/neu.2010.1597 1009

Susuki, K., Otani, Y., & Rasband, M. N. (2016). Submembranous cytoskeletons stabilize nodes 1010 of Ranvier. Experimental Neurology, 283, 446–451. 1011 https://doi.org/10.1016/J.EXPNEUROL.2015.11.012 1012

Tomassy, G. S., Berger, D. R., Chen, H.-H., Kasthuri, N., Hayworth, K. J., Vercelli, A., … 1013 Arlotta, P. (2014). Distinct Profiles of Myelin Distribution Along Single Axons of Pyramidal 1014 Neurons in the Neocortex. Science, 344(6181), 319–324. 1015 https://doi.org/10.1126/science.1249766 1016


https://doi.org/10.1101/2020.03.09.983569


31

Trapp, B. D., Nishiyama, A., Cheng, D., & Macklin, W. (1997). Differentiation and death of 1017 premyelinating oligodendrocytes in developing rodent brain. The Journal of Cell Biology, 1018 137(2), 459–468. https://doi.org/10.1083/jcb.137.2.459 1019

Vega-Riquer, J. M., Mendez-Victoriano, G., Morales-Luckie, R. A., & Gonzalez-Perez, O. 1020 (2019). Five Decades of Cuprizone, an Updated Model to Replicate Demyelinating 1021 Diseases. Current Neuropharmacology, 17(2), 129–141. 1022 https://doi.org/10.2174/1570159X15666170717120343 1023

Walhovd, K. B., Johansen-Berg, H., & Káradóttir, R. T. (2014). Unraveling the secrets of white 1024 matter--bridging the gap between cellular, animal and human imaging studies. 1025 Neuroscience, 276(100), 2–13. https://doi.org/10.1016/j.neuroscience.2014.06.058 1026

Werneburg, S., Jung, J., Kunjamma, R. B., Ha, S.-K., Luciano, N. J., Willis, C. M., … Schafer, 1027 D. P. (2020). Targeted Complement Inhibition at Synapses Prevents Microglial Synaptic 1028 Engulfment and Synapse Loss in Demyelinating Disease. Immunity, 52(1), 167-182.e7. 1029 https://doi.org/10.1016/j.immuni.2019.12.004 1030

Xiao, L., Ohayon, D., McKenzie, I. A., Sinclair-Wilson, A., Wright, J. L., Fudge, A. D., … 1031 Richardson, W. D. (2016). Rapid production of new oligodendrocytes is required in the 1032 earliest stages of motor-skill learning. Nature Neuroscience, 19(9), 1210–1217. 1033 https://doi.org/10.1038/nn.4351 1034

Yeung, M. S. Y., Djelloul, M., Steiner, E., Bernard, S., Salehpour, M., Possnert, G., … Frisén, J. 1035 (2019). Dynamics of oligodendrocyte generation in multiple sclerosis. Nature, 566(7745), 1036 538–542. https://doi.org/10.1038/s41586-018-0842-3 1037

Yeung, M. S. Y., Zdunek, S., Bergmann, O., Bernard, S., Salehpour, M., Alkass, K., … Frisén, J. 1038 (2014). Dynamics of oligodendrocyte generation and myelination in the human brain. Cell, 1039 159(4), 766–774. https://doi.org/10.1016/j.cell.2014.10.011 1040

Young, K. M., Psachoulia, K., Tripathi, R. B., Dunn, S.-J., Cossell, L., Attwell, D., … Richardson, 1041 W. D. (2013). Oligodendrocyte dynamics in the healthy adult CNS: evidence for myelin 1042 remodeling. Neuron, 77(5), 873–885. https://doi.org/10.1016/j.neuron.2013.01.006 1043

Zhang, Y., Zhang, J., Navrazhina, K., Argaw, A. T., Zameer, A., Gurfein, B. T., … John, G. R. 1044 (2010). TGFbeta1 induces Jagged1 expression in astrocytes via ALK5 and Smad3 and 1045 regulates the balance between oligodendrocyte progenitor proliferation and differentiation. 1046 Glia, 58(8), 964–974. https://doi.org/10.1002/glia.20978 1047

1048 1049 1050


https://doi.org/10.1101/2020.03.09.983569


32

ACKNOWLEDGEMENTS 1051 We thank Dr. M Pucak and N Ye for technical assistance, T Shelly for machining expertise, M 1052 Bhat for gift of the anti-Caspr antibody and members of the Bergles laboratory for discussions. J 1053 Orthmann-Murphy was supported by grants from the National Multiple Sclerosis Society and the 1054 Hilton Foundation. C Call and G Molina-Castro were supported by National Science Foundation 1055 Graduate Research Fellowships. Funding was provided by grants from the NIH (NS051509, 1056 NS050274, NS080153), a Collaborative Research Center Grant from the National Multiple 1057 Sclerosis Society, and the Dr. Miriam and Sheldon G Adelson Medical Research Foundation to 1058 D Bergles. 1059 1060 1061 Competing interests: 1062 JOM, CLC, GCM, YCH, MNR and DEB have no competing interests. PAC is PI on grants to 1063 JHU from Biogen and Annexon and has received consulting fees for serving on scientific 1064 advisory boards for Biogen and Disarm Therapeutics. 1065 1066


https://doi.org/10.1101/2020.03.09.983569


33

FIGURES AND LEGENDS 1067

1068


https://doi.org/10.1101/2020.03.09.983569


34

Figure 1: An in vivo platform to monitor loss and replacement of oligodendrocytes in the 1069 upper cortical layers. A: In vivo two photon microscopy through chronic cranial windows over 1070 the somatosensory cortex of Mobp-EGFP mice (coronal view), showing myelinated fibers in 1071 cortical layer I parallel to pial surface and in deeper layers oriented perpendicularly. B: Electron 1072 micrograph reconstruction of adult mouse visual cortex (from (Bock et al., 2011) illustrating low 1073 density of myelinated fibers (arrows) in the upper layers of cortex. C: Maximum intensity y-1074 projection (coronal view, 425 µm x 150 µm x 550 µm) and z-projection (top-down view, 425 µm 1075

x 425 µm x 100 µm) example regions from Mobp-EGFP mice with chronic cranial windows. D: 1076 Schematic illustrating longitudinal course of loss (demyelination) and replacement 1077 (remyelination) of cortical oligodendrocytes. E: Examples of maximum intensity projection 1078 images of the same region (156 µm x 156 µm x 84 µm) imaged repeatedly from an adult sham- 1079 (control, top row) or a cuprizone-treated (bottom row) mouse are shown with overlay of cell 1080 bodies from baseline (magenta) and after 6 weeks (green). Merge of baseline and 6 week 1081 overlays show where new cells are added to the region (arrows). F-G: Individual cells 1082 (represented by magenta, blue or green lines) were tracked longitudinally in somatosensory 1083 cortex from mice fed control (F; from region in top row of E) or cuprizone diet (G; from region in 1084 bottom row of E). H-K: The same cortical volume (425 µm x 425 µm x 300 µm) was imaged 1085 repeatedly in mice given either control or cuprizone diet, and individual cells present at baseline 1086 (black) or formed at later time points (green) were tracked over time. Shown are the average 1087 cell counts depicted as a proportion of baseline number of cells, (H, N = 5 control mice; I, N = 6 1088 cuprizone mice, I; number of mice imaged at each time point indicated). J-K: The average rate 1089 of loss (J) or addition (K) of oligodendrocytes per week in control-treated (blue) v. cuprizone-1090 treated mice (orange) relative to the baseline population of oligodendrocytes. Treatment with 1091 sham or cuprizone-supplemented chow denoted by shaded background. In cuprizone-treated 1092 mice, there was a higher rate of oligodendrocyte loss over weeks 3-5 (@ 3 weeks, p = 0.02014; 1093 @ 4 weeks, p=0.0000488; @ 5 weeks, p = 0.0066) and addition of new cells between 4-6 1094 weeks (@ 4 weeks, p = 0.0280; @ 5 weeks, p = 0.0121; @ 6 weeks, p = 0.000530) compared 1095 to control. Data is presented as means with standard error of the mean bars; data are compared 1096 with N-way ANOVA test with Bonferroni correction for multiple comparisons. 1097 1098


https://doi.org/10.1101/2020.03.09.983569


35

1099 Figure 2: Regeneration of oligodendrocytes is incomplete in lower cortical regions. A: 1100 The fate of individual oligodendrocytes were determined within the same cortical volume (425 1101 µm x 425 µm x 300 µm) over time that was divided into 0-100, 100-200, or 200-300 µm zones, 1102

(example maximum intensity Y projection is 425 µm x 150 µm x 300 µm). B-C: Mice were given 1103 either control diet (B) or cuprizone-supplemented diet (C), and individual cells present at 1104 baseline (black) or newly appearing over the time-course (green) were tracked longitudinally. 1105 The average cell counts per volume are depicted as a proportion of baseline number of cells (B: 1106


https://doi.org/10.1101/2020.03.09.983569


36

N = 5 control mice; C: N = 6 cuprizone-treated mice; with the same number of mice imaged at 1107 each time point as depicted in Fig. 1H,I). D-G: The rate of cell loss (D, F) or addition (E,G) 1108 relative to the baseline population of oligodendrocytes in each zone are depicted for each 1109 imaging time-point as a function of cortical depth (0-100 v. 200-300 µm zones), over 9 weeks of 1110 imaging in control (D-E) or cuprizone-treated (F-G) mice. Treatment with cuprizone-1111 supplemented chow denoted by shaded background (F-G). Cells are rarely lost in control 1112 regions (D), and the rate of cell addition in the 0-100 (light blue circles) v. 200-300 µm (dark blue 1113

squares) zones are similar (except @ week 5, p = 0.0036). In the bottom 200-300 µm zone (F-1114 G, in cuprizone-treated mice (brown squares), the rate of oligodendrocyte loss (F) is significantly 1115 greater at week 4 relative to the top (0-100 µm, orange circles) zone (p = 0.0443), and the rate 1116 of addition is significantly lower (G; p = 0.0364). H: The mean number of oligodendrocytes lost 1117

(magenta) and added (green) at each imaging time-point, comparing the 0-100 µm and 200-300 1118

µm zones. I: The distribution of baseline (gray) and new (green) oligodendrocyte cell bodies 1119

relative to the center of an example region volume (left-side panels; 425 µm x 425 µm x 300 1120

µm) and for all regions quantified (right-side panels; mean values from N = 6 mice) relative to 1121 the pial surface (0 on y-axis) is shown at baseline and at recovery weeks 1, 2 and 5 1122 (corresponding to week 4, 5, and 8 of imaging weeks from A-H). Mean values depicted with 1123 error bars as standard error of the mean, and statistical tests are N-way ANOVA with Bonferroni 1124 correction for multiple comparisons. 1125


https://doi.org/10.1101/2020.03.09.983569


37

1126 Figure 3: Remyelinating oligodendrocytes appear in novel locations but maintain 1127 morphology of oligodendrocytes generated in cuprizone-naïve mice. 1128 A-B: The location for all oligodendrocyte cell bodies present within the same cortical volume 1129

(425 µm x 425 µm x 300 µm) at baseline and after 8 weeks of two photon in vivo imaging are 1130 plotted and overlaid in 3-D. The fate of each cell over this time-course is designated as stable 1131 (black), lost (magenta), new (green). Example volumes for control (A) and cuprizone-treated (B) 1132 cortex are shown. C: The average 3-D nearest neighbor distance (NND) for lost, new, and 1133


https://doi.org/10.1101/2020.03.09.983569


38

stable cells, comparing baseline and last (8 weeks) imaging time-point locations, as plotted in A-1134 B, were determined. Stable cells in both control (blue) v. cuprizone-treated (orange) cortices are 1135 minimally displaced (schematized by two gray oligodendrocytes displaced by d distance) in both 1136 the full 425 µm x 425 µm x 300 µm (0-300 µm; p = 4.64) volumes and the surface 425 µm x 425 1137

µm x 100 µm (0-100 µm, p = 2.36) volume. New cells added into control cortex did not 1138

significantly alter NND in the full 0-300 µm region (controls cells (blue), all cells 0-300 µm v. 1139

stable cells 0-300 µm, p = 0.284), but did so in the top 0-100 µm zone, (controls cells (blue), all 1140

cells 0-100 µm v. stable cells 0-100 µm, p = 0.0137). The displacement of oligodendrocytes in 1141 cuprizone-treated cortex (schematized displacement (d) between lost (magenta) and new 1142 (green) cells) was greater in cuprizone-treated cortex (regenerated cells (orange), all cells 0-300 1143 µm v. stable cells 0-300 µm; p = 8.95 x 10-7), despite near complete replacement of cells in the 1144

top 0-100 µm zone (regenerated cells (orange), all cells 0-100 µm v. stable cells 0-100 µm; p = 1145 9.12 x 10-7). All comparisons in C are N-way ANOVA with Bonferroni correction for multiple 1146 comparisons. D: Every cell body (black), process (black) and myelin sheath (blue, control (10 1147 cells, N=3 mice); orange, cuprizone (9 cells, N=3 mice)) belonging to newly appearing 1148 oligodendrocytes in cortical layer I (top zone) were traced using Simple Neurite tracer (Image J) 1149 on day of appearance. Shown are examples of maximum intensity projections of these rendered 1150 pseudocolored tracings. E: Newly formed control and remyelinating oligodendrocytes have 1151 similar numbers of myelin sheaths (p = 0.628, unpaired two-tailed t-test). F-G: The total length 1152 of myelin formed by a new cell (F) and the average length per sheath (G) for new control (blue) 1153 and remyelinating (orange) cells. Remyelinating cells overall make significantly more myelin (F; 1154 p = 0.041, unpaired two-tailed t-test) and individual myelin sheaths are longer (G; p = 0.012, 1155 unpaired two-tailed t-test). H: The distribution of sheath lengths per cell traced is plotted (each 1156 column represents the distribution of sheath lengths for an individual cell at 12-14 days from 1157 appearance (control, blue; remyelinating, orange, with mean and SEM in gray). I-J: Vectors 1158 were calculated from the cell body extending to each paranode of a reconstructed 1159 oligodendrocyte at day 12-14, then x and y vector components were summed and oriented to 1160 same direction of the vector sum. These are plotted (I) in 2-D for each cell traced (Control, N = 1161 10 cells from 3 mice; Cuprizone, N = 9 cells from 3 mice), and were not significantly different 1162 from uniformly radial (J, Control, –0.047 ± 1.30 (std) rad, p = 0.400; Regenerated, 0.076 ± 1.30 1163 (std) rad, p = 0.256, Hodges-Ajne test of non-uniformity). The average circular morphologies for 1164 new control or remyelinating cells (J, same cells as in I) exhibit similar degrees of circularity (p > 1165 0.1, k = 462, Kuiper two-sample test). 1166 1167


https://doi.org/10.1101/2020.03.09.983569


39

1168

1169 Figure 4: Individual remyelinating oligodendrocytes myelinate cortical territory that is 1170 distinct from lost oligodendrocytes. 1171 A-B: The best fit ellipsoid (E) that encompasses 80% of all myelin sheaths of an individual 1172 oligodendrocyte (examples shown in A for new control (blue) and remyelinating (orange) cells in 1173

the surface 0 -100 µm zone (for x-y radius) and coronal (for z radius) views) was calculated and 1174 the average value ((Control (blue), N = 10 cells from 3 mice; Cuprizone (orange), N = 9 cells 1175 from 3 mice)) is plotted in B ((gray, cells at baseline (N = 7 cells from 4 mice)). Circles represent 1176 individual cells). Remyelinating cells were significantly wider (x-y radius) than new control cells 1177 (p = 0.025, one-way ANOVA). C-D: The average best-fit ellipsoids for control (C) or 1178 remyelinating cells (D) as calculated in (B) were plotted in example regions from the top 0-100 1179 µm of cortex (425 μm x 425 μm x 100 µm) based on location of cell bodies. E-F: The proportion 1180 of total territory volume at baseline that was overlapped by regenerated oligodendrocytes 1181 (scaled to take into account differences in numbers of baseline and regenerated cells; see 1182 Methods), and the additional volume encompassed by regenerated oligodendrocyte territory 1183


https://doi.org/10.1101/2020.03.09.983569


40

relative to total baseline volume (F), for 0-100 µm regions in cuprizone-treated (orange, N = 6 1184 mice) and compared to the volumes predicted if the same number of regenerated cells 1185 appeared at random within the same volumes (gray). Regenerated oligodendrocyte territories 1186 partially overlap with baseline volume (E; 59.1%, p = 0.0778 by one-way ANOVA) and 1187 encompass novel territory (F; 115%, p = 0.662 by one-way ANOVA), at a similar proportion to 1188 regenerated cells placed at random. 1189

1190


https://doi.org/10.1101/2020.03.09.983569


41

Figure 5: After oligodendrocyte loss, regeneration results in a new pattern of cortical 1191 myelin. A-B. Individual myelin sheaths which passed through a 100 μm x 100 μm x 100 μm 1192 volume within the top 0-100 μm zone were traced, and the fate of each sheath within the volume 1193 was determined to be stable (black), lost (magenta), replaced (blue) or novel (green) across the 1194 time series. Examples of traced myelin sheaths with pseudocolor designation from control (A) 1195 and cuprizone-treated (B) cortex are depicted. C: Total number of traced internodes at baseline 1196 and after 8 weeks of imaging are plotted (circles represent means for individual mice, with line 1197 connecting two time-points; mean for all mice is filled circle with SEM) for control (blue, N = 5), 1198 and cuprizone-treated (orange, N = 5) mice. D: Quantification of internode fates from cuprizone-1199 treated (orange, N = 5) and control (blue, N = 5) mice (circles represent mean proportional 1200 values relative to baseline from individual mice). Error bars are standard error of the mean. E: 1201 Baseline panel on left is a maximum intensity projection (226 μm x 226 μm x 60 μm) illustrating 1202 the myelin sheaths at baseline, with a red dashed line demarcating an area of higher (lower left) 1203 v. lower myelin density (top right). Right panel is a rendering of a completely traced and 1204 pseudocolored new oligodendrocyte that appears in the region at 3 days recovery and forms 1205 myelin sheaths that either replace those lost (magenta) or are novel (green). F: The proportion 1206 of replaced sheaths per new oligodendrocyte were plotted as a function of the number of 1207 baseline myelin sheaths present within the territory (average remyelinating ellipsoid, see Figure 1208 4A,B) of the new cells (13 remyelinating oligodendrocytes from N = 4 mice), correlation co-1209 efficient R2 = 0.4232. 1210


https://doi.org/10.1101/2020.03.09.983569


42

1211 1212 Figure 6: Myelin sheaths are more likely to be replaced if they had neighboring sheaths 1213 before demyelination. 1214 A: Schematic of intermittent myelination of cortical axons, designating each internode by 1215 number of flanking myelin sheath neighbors (0, yellow; 1 or 2, lavender; or undefined, gray). B: 1216 Example maximum intensity projections of lost (magenta) and new (green) myelin sheaths from 1217


https://doi.org/10.1101/2020.03.09.983569


43

cuprizone-treated mice. Top row is an example of a lost, but not replaced, sheath, as well as a 1218 novel sheath not present at baseline (green arrowhead). Remaining rows are examples of lost 1219 (magenta sheaths and arrowheads) and replaced internodes (green sheaths and arrowheads) 1220 that had 0, 1, or 2 neighboring sheaths at baseline. C, F: Comparison of mean proportion of 1221 internodes with at least one neighbor (lavender), isolated (yellow), or undefined (gray) within a 1222 100 μm x 100 μm x 100 μm volume at baseline and 8 weeks later, from control (C, N = 5) and 1223 cuprizone-treated (F, N = 5) mice. Volumes from both control and cuprizone-treated conditions 1224 have the same relative proportion of isolated vs. ≥ 1 neighboring internode at both time-points (p 1225 > 0.05, N-way ANOVA with Bonferroni correction for multiple comparisons). D, G: Comparison 1226 of the mean proportion of internodes with 0 or ≥ 1 neighbor that are stable or novel (control, D) 1227 or replaced, not replaced, or novel (cuprizone, G). There is no significant difference in the 1228 proportion of isolated vs. ≥ 1 neighbor population between stable and novel sheaths in control 1229 (D, p > 0.05, unpaired two-tailed t-test with Bonferroni correction for multiple comparisons), nor 1230 between lost and novel sheaths in cuprizone (G, p > 0.05, unpaired two-tailed t-test with 1231 Bonferroni correction for multiple comparisons), but relatively more internodes with ≥ 1 neighbor 1232 were replaced in cuprizone-treated cortex (G, p = 5.93 x 10-7, unpaired two-tailed t-test with 1233 Bonferroni correction for multiple comparisons). E,H: The average number of internodes 1234 categorized by number of neighbors at baseline and then at final imaging time-point, plotted in a 1235 “myelination matrix” for control (E), where, there is a trend to increase number of neighbors over 1236 8 weeks. In cuprizone-treated mice, the average number of neighbors for lost and replaced 1237 internodes are categorized by number of neighbors at baseline (lost internode) and after 5 1238 weeks of recovery (8 weeks imaging, replaced internode). Internodes with more neighbors at 1239 baseline are more likely be replaced (largest average # of internodes in bottom right of matrix). 1240


https://doi.org/10.1101/2020.03.09.983569


44

1241 Figure 7: Structural components of the node of Ranvier persist after demyelination. 1242 A: Adult Mobp-EGFP mice were fed cuprizone-supplemented diet for three weeks and individual 1243 myelin sheaths were imaged and traced within a 100 μm x 100 μm x 100 μm volume to 1244


https://doi.org/10.1101/2020.03.09.983569


45

determine their fate. An example of myelin sheaths that were lost (magenta, baseline), and 1245 regenerated (green, 5 weeks recovery) are overlaid on maximum intensity projections from a 1246 longitudinally imaged somatosensory cortical region. These traced myelin sheaths are shown 1247 degenerating at 5 days recovery. Orange arrowhead denotes location of node of Ranvier at 1248 baseline and a similar position after 2 neighboring myelin sheaths are regenerated at 5 weeks of 1249

recovery. B. Schematic depicting axonal regions of a myelinated axon where βIV-spectrin (node 1250

of Ranvier, red) and Caspr (paranode, cyan) localize. After demyelination, puncta of βIV-1251 spectrin can be found with 2 flanking Caspr puncta (“Nodes”), 1 flanking Caspr punctum 1252 (“Heminode”), or no nearby Caspr puncta (“Isolated”). C-D: Adult Mobp-EGFP mice were fed 1253 cuprizone-supplemented diet (D) or sham chow (C) for 6 weeks to induce loss of 1254 oligodendrocytes and inhibit formation of new oligodendrocytes. Coronal sections from brains of 1255 mice euthanized after 4 or 6 weeks of treatment were immunostained for βIV-spectrin and 1256 Caspr. E-F: Magnified views of immunostained somatosensory cortex (Raw data) from control 1257 (E) and cuprizone-treated (F) brains. Example post-processed “Surfaces” that were used to 1258

calculate nearest neighbor distances between βIV-spectrin puncta and Caspr puncta within 3.5 1259

µm are shown for the same magnified images in E and F. Axon initial segments were excluded 1260

from surface rendering. G: Total number of βIV-spectrin puncta categorized as either Node, 1261

Heminode or Isolated. There are fewer overall βIV -spectrin puncta after 6 weeks of cuprizone 1262 (compared to controls (@4 or 6 weeks) or 4 weeks of cuprizone, but relatively more isolated 1263 puncta still present after 6 weeks of cuprizone treatment. (control @ 4 weeks, N = 3 mice; 1264 control @ 6 weeks, N = 3 mice; cuprizone @ 4 weeks, N = 4 mice; cuprizone @ 6 weeks, N = 4 1265 mice). Bars are SEM. 1266 1267 1268 1269


https://doi.org/10.1101/2020.03.09.983569


46

SUPPLEMENTARY DATA 1270 1271

1272 Supplementary Figure 1: Degeneration of oligodendrocytes in cuprizone-treated Mobp-1273 EGFP mice. Shown are two examples of individual oligodendrocytes tracked longitudinally 1274 using two photon in vivo imaging through chronic cranial windows in cuprizone-fed adult Mobp-1275 EGFP mice. A: Example of an oligodendrocyte present at baseline (cell body denoted with 1276 magenta arrowhead) that loses GFP signal in processes and myelin sheaths and eventually cell 1277 body by 3 weeks of cuprizone treatment (maximum intensity projection of 156 µm x 156 µm x 45 1278

µm volume). B. Example of an oligodendrocyte present at baseline that loses GFP signal in 1279 processes and myelin sheaths over a much longer time course than the cell in a, eventually 1280 disappearing at 3 weeks of recovery, after new oligodendrocytes (green arrowheads) are 1281

formed during recovery period (maximum intensity projection of 156 µm x 156 µm x 55 µm 1282 volume). 1283 1284


https://doi.org/10.1101/2020.03.09.983569


47

1285 Supplementary Figure 2: Myelin is not uniformly distributed throughout the adult rodent 1286 cortex. A-C: The left hemi-cortex from an adult Mobp-EGFP mouse is flattened and 1287 immunostained for EGFP (A), ASPA (B, oligodendrocytes) and MBP (C, myelin), merged 1288 together (E) and individual regions (schematized map in D) are expanded to illustrate that some 1289 cortical areas have a higher density of MBP+ myelin sheaths and GFP+/ASPA+ 1290 oligodendrocytes (top, primary somatomotor cortex) than others (bottom, auditory cortex). F: 1291 Coronal section from a 6-month old wild-type mouse immunostained for MBP shows regional 1292 heterogeneity of myelin across cortical mantle. 1293 1294 1295


https://doi.org/10.1101/2020.03.09.983569


48

1296


https://doi.org/10.1101/2020.03.09.983569


49

Supplementary Figure 3: Distribution of astrocytes and oligodendrocyte precursor cells 1297 over the course of cuprizone treatment and recovery. A: Example coronal images from the 1298 brains of young adult Mobp-EGFP mice euthanized at baseline or following 3 weeks of 1299 cuprizone administration followed by 1, 2, 3 or 5 weeks recovery. Sections were immunostained 1300 for EGFP (oligodendrocytes), NG2 (OPCs) and GFAP (astrocytes). By 2 weeks of recovery, the 1301 relatively sparse distribution of EGFP+ cells represent newly formed cells (as demonstrated by 1302 in vivo imaging in Figure 1), with increasing number of new EGFP+ cells in lower cortical layers 1303 in later weeks of recovery. NG2+ OPC distribution remains constant over the course of damage 1304 and repair. GFAP+ cells increase in number after cuprizone and remain elevated in lower 1305 cortical regions over several weeks of recovery. B: Example maximum intensity projections of 1306 coronal sections from Mobp-EGFP mice euthanized at baseline, 2 weeks recovery or 5 weeks 1307 recovery, immunostained for GFAP (213 µm x 213 µm x 35 µm, from A). GFAP+ astrocyte 1308 morphology becomes reactive after exposure to cuprizone and maintains reactive morphology, 1309 even at 5 weeks of recovery. C: Example of a baseline somatosensory cortex coronal section 1310 from an Mobp-EGFP mouse, immunostained for NG2+, and divided into 100 µm zones from the 1311

pial surface to 500 µm in depth. The cell body location is marked with a circle. Each 100-µm 1312 zone color coded from pial surface corresponds bar colors in D and E. D-E: Quantification of 1313 cortical NG2+ cell distribution (D) , and GFAP+ astrocytes (E) from brains of adult Mobp-1314 EGFP mice euthanized at baseline (N = 4), 1 week of cuprizone (N = 4), 2 weeks of cuprizone 1315 (N = 4), 3 weeks of cuprizone (N = 4), 1 week of recovery (N = 8), 2 weeks of recovery (N = 5), 1316 or 3 weeks of recovery (N = 5 for NG2, N = 4 for GFAP), 5 weeks of recovery (N = 4). F-G. The 1317 ratio of cell number in the last (400-500 µm, green in C) versus first (0-100 µm, blue in C) zone 1318 for NG2+ OPCs (F) and GFAP+ astrocytes (G). Compared to baseline, the relative proportion of 1319 NG2 cells in top vs. bottom regions is stable over the course of cuprizone-treatment and 1320 recovery (F; p = 0.0858; Kruskal-Wallis one-way ANOVA with Fisher’s LSD) whereas the 1321 number of GFAP+ cells significantly increase in the bottom zone after 2 weeks of cuprizone and 1322 remain elevated (G, p = 0.0056, Kruskal-Wallis one-way ANOVA with Fisher’s LSD ). Error bars 1323 are standard error of the mean. 1324 1325


https://doi.org/10.1101/2020.03.09.983569


50

1326 1327 Supplementary Figure 4: Dynamics of oligodendrocyte maturation in adult Mobp-EGFP 1328 mice. Every cell body, and associated myelin sheath belonging to newly appearing 1329 oligodendrocytes in cortical layer I were traced using Simple Neurite Tracer (Image J) on day of 1330 appearance and imaged every 1-3 days for up to 14 days. A: An individual myelin sheath 1331 (demarcated by blue arrowheads at the paranodal tips) at day 0 was followed over 14 days. The 1332 left-side paranode extends (through day 5) and then retracts and the right-side paranode 1333 extends to encounter a neighboring sheath and subsequently flanks a node of Ranvier (day 14). 1334 The yellow arrow marks the position where the oligodendrocyte process connects to the myelin 1335


https://doi.org/10.1101/2020.03.09.983569


51

sheath. B: The overall length of the individual sheath shown in A increases over 14 days. C: 1336 Example maximum intensity projections of traced processes and myelin sheaths from newly 1337 formed cells in control or cuprizone-treated mice on day of appearance, that were followed over 1338 14 days and length of individual traced sheaths are denoted as either unchanged (black) or 1339 exhibiting net extension (green) or retraction (magenta). D: There were significantly more myelin 1340 sheaths undergoing extension than retraction in newly formed cells (control, p = 4.11 x 10-7; 1341 remyelinating, p = 0.00119, unpaired two-tailed t-tests, with Bonferroni correction for multiple 1342 comparisons) (top) but no difference in net length change of extensions or retractions (bottom), 1343 and no significant difference between control or remyelinating cells. E-F: There were low 1344 numbers of myelin sheaths lost in newly formed oligodendrocytes (E) and cytoplasmic 1345 processes were retracted (F); but there were no significant differences between control or 1346 remyelinating cells (sheaths, p = 0.907, unpaired two-tailed t-test; processes, p = 0.474, 1347 unpaired two-tailed t-test; control N = 10 cells, remyelinating N = 10 cells)). G-H: Examples of 1348 cytoplasmic processes and myelin sheaths (cyan) present at day 0 in a newly formed 1349 remyelinating cell, that are no longer present at day 7 (G) or day 4 (H). In H, the myelin sheath 1350 is dissolved first (day 2) and then the process connecting it to the cell body retracts completely 1351 by day 4. I-J: The absolute value of net total change in myelin sheath length over time is plotted 1352 for newly formed traced cells in control (blue) and cuprizone-treated (orange) cortex (thick line 1353 depicts means, thin lines represent individual cells). The majority of the length changes occur in 1354 the first 4 days after appearance (p = 2.28 x 10-6, N-way ANOVA with Bonferroni correction for 1355 multiple comparisons). J: The summed total length of all myelin sheaths per newly formed cell 1356 are plotted as a proportion of the total length at day of appearance. The overall trend is 1357 extension of myelin for both control and remyelinating cells. K: The length of individual myelin 1358 sheaths across all reconstructed cells plotted against the contact point of the cytoplasmic 1359 process to the myelin sheath, where 0 represents the center of an individual sheath (example in 1360 L, top panel, cyan process intersecting the cyan sheath towards the center of the sheath at 1361 orange arrow) and 1 or -1 are the distal tips of the sheath (example in L, bottom panel, cyan 1362 process intersecting at the paranode of the cyan sheath at orange arrow). 1363 1364 1365 Supplementary Video 1: Loss and replacement of oligodendrocytes. Longitudinal in vivo 1366

imaging of demyelination and remyelination. This is a 392 µm x 392 µm x 100 µm volume 1367 shown as a maximum intensity projection that was repeatedly imaged through a chronic cranial 1368


https://doi.org/10.1101/2020.03.09.983569


52

window over the somatosensory cortex in an adult Mobp-EGFP mouse, at baseline, over 3 1369 weeks of cuprizone administration, and then through 5 weeks of recovery. Scale bar is 50 µm. 1370 Link: 1371 https://drive.google.com/file/d/18dTdmKwywMHbMt81p_64AshmRvgD8OEX/view?usp=sh1372 aring 1373 1374 Supplementary Video 2: New oligodendrocytes are added in the upper cortical layers in 1375 adult mice. Longitudinal imaging of an adult Mobp-EGFP mouse with a chronic cranial window 1376

fed sham diet. Region corresponds to images shown in Figure 1E, top row. Scale bar is 25 µm. 1377 Link: 1378 https://drive.google.com/file/d/1b6up2_RWkh5qetrvYIqGv7kBQirTkFSi/view?usp=sharing 1379 1380 Supplementary Video 3: Oligodendrocytes are lost and new cells appear after cuprizone-1381 treatment. Longitudinal imaging of an adult Mobp-EGFP mouse with a chronic cranial window 1382 fed 3 weeks of a cuprizone-supplemented diet followed through 5 weeks of recovery. Region 1383 corresponds to images shown in Figure 1E, bottom row. Scale bar is 25 µm. 1384 Link: 1385 https://drive.google.com/file/d/1keWRJVo8kbN1khxA0tzdQzZV8GsPrL0j/view?usp=sharin1386 g 1387 1388 Supplementary Video 4: Myelin sheaths are lost and novel sheaths are formed after 1389 cuprizone-treatment. Longitudinal imaging of an adult Mobp-EGFP mouse with a chronic 1390 cranial window fed 3 weeks of a cuprizone-supplemented diet followed through 5 weeks of 1391 recovery. A myelin sheath at baseline (traced and pseudocolored magenta, overlayed in 1392 maximum intensity projection of longitudinally-imaged region), degenerates over time (only the 1393 traced sheath from baseline is shown in subsequent time-points, and is lost by 1 week of 1394 recovery). At 5 days of recovery a novel isolated sheath (not present at baseline, traced and 1395 pseudocolored in green in the 5 week recovery time-point overlay) appears, formed by a 1396 remyelinating oligodendrocyte not present at baseline (cell in 5 week recovery time-point 1397 overlay). Scale bar is 15 µm. 1398 Link: 1399 https://drive.google.com/file/d/1fyVMqlAPOTH45SlOfgpWOg_yC-1400 Pol45U/view?usp=sharing 1401 1402


https://doi.org/10.1101/2020.03.09.983569


53

Supplementary Video 5: Isolated myelin sheaths are replaced. Longitudinal imaging of an 1403 adult Mobp-EGFP mouse with a chronic cranial window fed 3 weeks of a cuprizone-1404 supplemented diet followed through 5 weeks of recovery. An isolated myelin sheath at baseline 1405 (traced and pseudocolored magenta, overlayed in maximum intensity projection of 1406 longitudinally-imaged region), degenerates over time (only the traced sheath from baseline is 1407 shown in subsequent time-points, and is lost by 3 days of recovery). At 5 days of recovery a 1408 replacement isolated sheath appears (traced and pseudocolored in the 5 week recovery time-1409 point overlay), formed by a remyelinating oligodendrocyte not present at baseline (cell in 5 week 1410 recovery time-point overlay). Scale bar is 15 µm. 1411 Link: 1412 https://drive.google.com/file/d/1T6-Y0-1413 rz7Gx32zjJMSaLcq9B6AAYETvM/view?usp=sharing 1414 1415 Supplementary Video 6: Neighboring myelin sheaths with at least one neighbor are 1416 replaced and form a node of Ranvier in close proximity to one present at baseline. 1417 Longitudinal imaging of an adult Mobp-EGFP mouse with a chronic cranial window fed 3 weeks 1418 of a cuprizone-supplemented diet followed through 5 weeks of recovery. Two neighboring 1419 myelin sheaths (pseudocolored magenta in baseline time-point), flank an unlabeled node of 1420 Ranvier (paranodal loops of myelin accumulate cytoplasmic EGFP in Mopb-EGFP mice 1421 (Hughes et al., 2018)); these sheaths were traced over each imaging time-points, and 1422 degenerate after cuprizone-treatment. Remyelinating oligodendrocytes (one shown in 5 week 1423 recovery time-point) form replacement myelin sheaths (pseudocolored green in 5 week recovery 1424 time-point); appear to flank an unlabeled node of Ranvier in a similar position as baseline. Scale 1425 bar is 15 µm. 1426 Link: 1427 https://drive.google.com/file/d/1jA_lK-Cj3_TGlNfeJ32n-6GQ8E2OeLSv/view?usp=sharing 1428 1429 Supplementary Table: Key Resources 1430 Primary antibodies 1431 Target Protein/markers

Host species

Source Dilution Catalog # Identifier

Aspartoacylase

(ASPA)

Rabbit Genetex 1:1500 GTX113389 RRID:AB_2036283


https://doi.org/10.1101/2020.03.09.983569


54

GFP Goat Cell

Signaling

1:500 5664 RRID:AB_10705523

GFP Chicken Aves Lab 1:1500 GFP-1020 RRID:AB_2307313

GFP Rabbit Richard

Huganir Lab

1:1000 JH40330 Gift from R. Huganir

MBP Mouse Sternberger 1:2000 808401 RRID:AB_2564741

MBP Chicken Aves Lab 1:500 F-1005 RRID:AB_2313550

NG2 Guinea

pig

Generated in

D.E. Bergles

lab against

entire NG2

protein

1:10,000 n/a Kang et al., 2013;

PMID: 23542689

GFAP Rabbit Dako 1:500 N1506 RRID:AB_10013482

Beta IV Spectrin Rabbit Generated by

M. Rasband

Lab

1:300 Gift from M. Rasband

lab

Beta IV Spectrin Chicken Generated by

M. Rasband

Lab


Ankrin G Rabbit Generated by

M. Rasband

Lab


lab

Caspr Guinea

pig

Generated by

Manzhoor

Bhat Lab

(Department

of Cellular

and

Integrative

Physiology

UT Health

Science

1:1500 Gift from M. Bhat


https://doi.org/10.1101/2020.03.09.983569


55

Center San

Antonio)

1432 Secondary antibodies 1433 Target Species

Conjugate Source Dilution Catalog # Identifier

Anti-Rabbit Alexa 488 Jackson

Immuno 1:2000

711-546-

152 RRID:AB_2340619

Anti-Rabbit Cy3 Jackson

Immuno 1:2000

711-166-

152 RRID:AB_2313568

Anti-Rabbit Cy5 Jackson

Immuno 1:2000

711-606-

152 RRID:AB_2340625

Anti-Mouse Alexa 488 Jackson

Immuno 1:2000

715-546-

150 RRID:AB_2340849

Anti-Mouse Cy3 Jackson

Immuno 1:2000

715-166-

151 RRID:AB_2340817

Anti-Mouse Cy5 Jackson

Immuno 1:2000

715-175-

151 RRID:AB_2340820

Anti-Guinea

Pig FITC

Jackson

Immuno 1:2000

706-096-

148 RRID:AB_2340454

Anti-Guinea

Pig Cy3

Jackson

Immuno 1:2000

706-166-

148 RRID:AB_2340461

Anti-Guinea

Pig Cy5

Jackson

Immuno 1:2000

706-606-

148 RRID:AB_2340477

Anti-Goat Alexa 488 Jackson

Immuno 1:2000

705-546-

147 RRID:AB_2340430

Anti-Goat Cy3 Jackson

Immuno 1:2000

705-166-

147 RRID:AB_2340413

Anti-Chicken Alexa 488 Jackson

Immuno 1:2000

703-546-

155 RRID:AB_2340376

Anti-Chicken Cy5 Jackson

Immuno 1:2000

703-006-

155 RRID:AB_2340347

1434


https://doi.org/10.1101/2020.03.09.983569


56

Software and Algorithms 1435 Name Source Identifier ZEN Blue/Black Zeiss RRID:SCR_013672

ImageJ https://imagej.nih.gov/ij/ RRID:SCR_003070

Fiji http://fiji.sc RRID:SCR_002285

Adobe Illustrator CS6 Adobe RRID:SCR_014198

MATLAB Mathworks RRID:SCR_001622

SyGlass IstoVisio RRID:SCR_017961

Imaris Bitplane RRID:SCR_007370

1436


https://doi.org/10.1101/2020.03.09.983569


remyelination alters the pattern of myelin in the cerebral cortex · cerebral cortex, we performed...

Documents